USER MANUAL - Racurs13.1.6.Controloferrorsoftiepointsmeasurements.....89 13.1.7.Controloferrorsofautomatictiepointsmeasurements.....90 1.Purposeofthedocument

Digital Photogrammetric System

Version 7.0

USER MANUALBlock adjustment

Table of Contents1. Purpose of the document .......................................................................................................... 42. General information .................................................................................................................. 43. The “Block adjustment” toolbar .................................................................................................. 54. Adjustment batch mode ............................................................................................................ 65. Objects displaying settings ........................................................................................................ 7

5.1. Parameters of points and error display on a block scheme ............................................... 75.2. Setup of points symbols ................................................................................................. 95.3. Points filter .................................................................................................................. 115.4. Points selection ........................................................................................................... 145.5. The “Point attributes” window ........................................................................................ 165.6. List of triangulation points ............................................................................................. 19

6. Brief residuals report ............................................................................................................... 227. Adjustment of central projection images block .......................................................................... 23

7.1. Workflow during adjustment of central projection images block ....................................... 237.2. The “Coordinate system” tab ........................................................................................ 247.3. The “Adjustment” tab .................................................................................................... 24

7.3.1. Adjustment parameters of central projection images block ................................... 247.3.2. Adjustment in free model ................................................................................... 277.3.3. Independent strips method ................................................................................ 287.3.4. Method of initial approximation by a block scheme .............................................. 307.3.5. Independent stereopairs method ........................................................................ 327.3.6. Bundle adjustment method ................................................................................ 347.3.7. Bundle adjustment with self-calibration ............................................................... 367.3.8. Parameters of camera self-calibration ................................................................. 387.3.9. Compensation of systematic errors .................................................................... 427.3.10. Select subblock ............................................................................................... 44

7.4. The “Report” tab .......................................................................................................... 458. Adjustment of scanner blocks .................................................................................................. 49

8.1. Workflow of scanner block adjustment ........................................................................... 498.2. The “Points” tab ........................................................................................................... 518.3. The “Images” tab ......................................................................................................... 54

8.3.1. Parameters of scanner images adjustment ......................................................... 548.3.2. Rigorous method of adjustment .......................................................................... 578.3.3. RPC method ..................................................................................................... 598.3.4. Generic adjustment method ............................................................................... 608.3.5. Manual creation of stereopairs ........................................................................... 628.3.6. Groups of images acquired from the same orbit .................................................. 66

8.4. The “Report” tab .......................................................................................................... 708.5. Recommendations on monoblock adjustment ................................................................ 738.6. Number of points recommended for adjustment of scanner block .................................... 77

9. Adjustment procedure ............................................................................................................. 7910. Creating adjustment report .................................................................................................... 8011. Adjustment accuracy control .................................................................................................. 8212. Data export and import .......................................................................................................... 83

12.1. Export of initial data .................................................................................................... 8312.2. Import of adjustment results ........................................................................................ 85

13. Stages of accuracy control .................................................................................................... 8813.1. Adjustment control in free model ................................................................................. 88

13.1.1. Splitting on subblocks ...................................................................................... 8813.1.2. Blunders control in camera focal length ............................................................ 8813.1.3. Blunders control in strips tying .......................................................................... 8913.1.4. Intermediate control of tie measurements blunders ............................................ 8913.1.5. Blunders detection on interstrip not-transferred points ........................................ 89

2

Block adjustmentPHOTOMOD 7.0

13.1.6. Control of errors of tie points measurements ..................................................... 9013.1.7. Control of errors of automatic tie points measurements ...................................... 90

13.2. Adjustment control with geodetic reference .................................................................. 9113.2.1. Control of coordinate system correctness ......................................................... 9113.2.2. Control of adjustment performed using projection centers coordinates ................ 9213.2.3. Control of adjustment performed without projection centers ................................ 9213.2.4. Control of bundle adjustment errors .................................................................. 93

14. Special aspects of airborne blocks adjustment in the system ................................................... 9314.1. Use the coordinates of projection centers .................................................................... 9314.2. Survey with big images overlap ................................................................................... 9414.3. Survey by long-focus cameras .................................................................................... 9514.4. Survey in different seasons ......................................................................................... 9514.5. Survey of forested areas ............................................................................................. 9614.6. Bundle adjustment ..................................................................................................... 9614.7. Tie points measurement on lengthy objects boundaries ................................................ 96

Appendix A. Exterior orientation angular systems ......................................................................... 96

3


1. Purpose of the documentThis document contains detailed information about block adjustment procedure inPHOTOMOD system’s window. Here there is a description of project types and methodsof block adjustment of both central projection images and scanner images, distinguishingfeatures of the methods, ways of adjustment accuracy control. The document also de-scribes special aspects of adjustment of different images blocks and accuracy controlof various types of projects on stage of projects processing in.

2. General informationBlock adjustment could be performed of both central projection images and scannerspace images.

The adjustment is a process of calculation exterior orientation parameters and coordinatemeasurement of all points in the block.

It is necessary to have interior and relative orientation of block images available to start blockadjustment.

Block adjustment is performed after stage of data collection for phototriangulation, in-ternal and relative images orientation, measurement of ground control points (GCP)(see “Aerial triangulation” User Manual).

Minimal requirement for block adjustment is available results of interior and relativeorientation of block images.

For adjustment of images block it is necessary to preliminary setup of adjustmentparameters – specify coordinate system, select adjustment method, choose main ad-justment parameters, setup report parameters and define residuals threshold of adjust-ment.

The system provides the following adjustment methods for adjustment of central projec-tion images block:

• Initial approximation independent strips method or initial approximation using blockscheme is basically used to reveal blunders, such as incorrectly specified groundcontrol points coordinates, errors of tie points positioning, etc.;

• independent stereopairs method is used to increase accuracy of adjustment results,obtained during initial approximation calculation; to recognize small errors and incase of acceptable results, for final adjustment;

• Bundle adjustment is used for final block adjustment;

In many cases it is advisable to use independent stereopairs method and bundle adjustmentalternatively to recognize and fix small errors.

4


measurement.pdf

• Freemodel – block adjustment without georeference data. When freemodel is createdit is possible to pass directly to project processing (vectorization, TIN and DEM cre-ation). Therefore, when GCP coordinates will be obtained the main work with projectwill be completed.

After interior and relative orientation it is necessary to measure GCP coordinates andadjust a block in selected coordinate system. After that all data obtained during pro-cessing project (vector objects, TIN and DEM) created in free model, will be transformedin selected coordinate system.

The system provides the following adjustment methods for adjustment of scanner imagesblock:

• Rigorous method – the method allows to consider exterior orientation parametersfrom metadata obtained from RS data provider;

• RPC-method – the method uses RPC-coefficients from metadata obtained from RSdata provider;

• Generic method – the method uses parallel-perspective model or Direct LinearTransformation (DLT) algorithm, that allows to process any scanner images (includingIRS, LandSat, etc.). However this requires more ground control points per stereopairas compared to rigorous method;

• Import of adjustment – import of images exterior orientation parameters frommetadatafor ADS 40/80/100 projects.

After images block adjustment the system provides possibility to view adjustment reportfor adjustment accuracy control, to edit points position for adjustment errors recovery,to save and export adjustment results of block scheme. Besides, the system allows toimport adjustment results, acquired in third party software packages for further projectprocessing using these data.

3. The “Block adjustment” toolbarThe Block adjustment additional toolbar is used for adjustment of images block.

To show the Block adjustment toolbar choose the Orientation › Block adjustment(Ctrl+Alt+S) or click the button on the main system toolbar.

Table 1. The “Block adjustment” toolbar

FunctionButtonsallows to load or restore current status of a block(load adjustment results after reloading project)allows to open window used to setup parametersof block adjustment (see Section 7 and Section 8)

5


FunctionButtonsallows to start the adjustment procedureallows to open window with results of block adjust-ment (see Section 10)allows to save results of block adjustment in orderto pass to the next stage of project processing inthe systemallows to reset block to initial state[only for central projection] allows to export adjust-ment results for further adjustment in third partysoftware[only for central projection] allows to import adjust-ment results obtained using third party softwareallows to display a window of displaying triangulationpoints settingsallows to show the Triangulation pointswindow witha list of all triangulation points of a project (groundcontrol, check, and tie) and also to view, search,edit, and import/export pointsallows to show/hide the window with values of RMS,mean absolute and maximal adjustment residuals.The window is intended for quick results estimationwithout viewing of detailed report (see Section 6)allows to show/hide the Point attributes windowwith information about selected point and possibilityto manage a type of points (see Section 5.5)allows to select on a block scheme types of pointschosen in the window (see Section 5.5)

4. Adjustment batch modeThe system also provides block adjustment in batch mode. Perform the following actionsto do this:

1. Start the server/client of distributed processing mode (see the “Distributed pro-cessing” chapter in the “General information” User Manual).

2. Choose Orientation › Adjustment in batch mode or click the button on the ATadditional toolbar. The Parameters window opens.

3. Setup parameters of central projection or scanner image block adjustment.

4. Click OK. Distributed processing tasks are created and the system shows amessageabout number of created tasks.

6


general.pdf

5. Objects displaying settings

5.1. Parameters of points and error display on a block scheme

The system provides setup of display of triangulation points on the block scheme anddisplay errors on points.

To setup objects display is used the Display options window. Click the button onthe Adjustment toolbar to open the window.

Fig. 1. Parameters of triangulation points display

The Display options window contains the toolbar with buttons used to perform thefollowing operations:

• – allows to display a list of points symbols on a block scheme;

• – allows to set on/off the filter of triangulation points display;

• – allows to setup the filter of triangulation points display;

In the Mode: field is displayed chosen mode of block scheme, it depends on active 2D-window – Block scheme or Image.

Triangulation points are not displayed in the stereopair 2D-window.

The Point display settings section allows to setup the following parameters:

• The Show method of points:

○ symbols – allows to display all points by symbols;

7


○ points – allows to display by points and also to specify the point size in pixels.

• Show names the following points in 2D-window:

○ All points;

○ Selected points;

○ Don’t show points names on a block scheme.

The Show errors section allows to choose errors of different types to display in 2D-window:

• By ground XY – vector with beginning in point’s center, directed to specified point’sposition, with small circle at the end;

• By ground Z – vertical vector, with small horizontal bar at the end.

Vectors mentioned above are displayed by green color, if residuals are within specifiedthreshold. If the residuals exceed the threshold, they are shown by red color.

• By ties between stereopairs XY – sidelong cross with size proportional to residualsvalues by X and Y;

• By ties between stereopairs Z – vertical line, directed symmetrically up and downin relation to point; the line’s size corresponds to residuals values by Z;

• By ties on images – circle with center in point’s center and with size proportional toresiduals values.

Vectors mentioned above are displayed by dark green color, if residuals are within specifiedthreshold. If the residuals exceed the threshold, they are shown by dark red color.

Residuals vectors with by ties types are not displayed in single image 2D-window.

Residuals vectors displaying (depending on its size) also could be changed in the Auxiliarysection of the Select points window (see the Section 5.3);

The On excluded points checkbox allows to show residuals for points excluded fromadjustment.

View scale section allows to manage zoom of residuals vectors display on a blockscheme in accordance with selected scale option:

• real scale – residual vector size (size of circle with a cross) corresponds to a schemescale, i.e. for residuals on GCP/check points the end of error vector indicates pointwith user defined coordinates at zoom “x1” (one to one) , that allows to see position

8


of erroneously recognized GCP. At that vectors are zoom in/out during zooming in/outof block scheme (see and Section 5.2);

• screen scale – in the field is used to specify number of units of coordinate system(meters, feet) in 100 screen pixels.

5.2. Setup of points symbols

Unique symbol and different colors can be used for each type of points on a blockscheme.

The Symbols window allows to display symbols list, change symbol or color. To openthe window click the button of the Display options window toolbar.

Fig. 2. The “Symbols” window

For points with known coordinates the following symbols are used:

• points with known X,Y coordinates – rectangle ;

• points with known Z coordinates – circle ;

• points with known X,Y,Z coordinates – triangle program.

For points with known coordinates the following symbol colors are used:

9


• for GCP – red color ;

• for check points – black color ;

• for points excluded from adjustment (C) – white color .

By default for tie strip points there are the following shapes and colors of symbols:

• single point – black circle;

• measured on a single stereopair – yellow rhombus;

• measured in triple or more overlap– green rhombus;

• measured on non-adjacent images – red circle;

• measured on a stereopair and its non-adjacent images – red circle.

By default for tie interstrip points there are the following shapes and colors of symbols:

• measured on single images in strips – crimson rhombus;

• measured on stereopairs – turquoise rhombus;

• measured on stereopairs and triplets – turquoise rhombus;

• measured on stereopairs and single images – crimson rhombus;

• measured on stereopairs and non-adjacent images – red circle;

• measured on non-adjacent images – red circle.

For the rest points there are the following shapes and colors of symbols:

• excluded – white square brackets ;

• without adjustment – red square brackets;

• targeted points – black upturned triangle ;

• projection centers – black rhombus .

A point could be displayed on a block scheme as combination of one, two or more symbols.

To change shape and color of tie points symbol (both strip and interstrip) perform thefollowing actions:

10


1. [optional] Click the button on theBlock adjustment toolbar to show theDisplayoptions window.

2. Click the button of the Display options window toolbar. The Symbols windowis opened.

3. In the By ties list select a point type which symbol or color needed to be changed.

4. Select Symbol to define a point’s shape – circle , rhombus or square .

5. Click the Color field and select necessary color from standard color palette ofWindows.

6. [optional] To go back to default marker size use the Reset button.

7. Click OK to confirm changes.

5.3. Points filter

It is possible to choose points for display on the block scheme. The Point filter windowis used for that.

To open the Point filter window, in the Display options window click the button.

11


Fig. 3. The “Point filter” window

To display points on a block scheme special – symbols.

The Select points window contains the following sections which are used to makepoints selection using different criteria:

• the By coordinates section is used to select points depending on their coordinates:

○ ground control points (ground and projection centers);

○ check points (ground and projection centers);

○ excluded – points with known coordinates, excluded from adjustment;

○ no coordinates – points without geodetic coordinates (if the checkbox is set off,only ground control points are displayed);

○ all – allows to set on/off all checkboxes of section.

12


XYZ,XY and Z points checkboxes allows to choote type of displayed ground control,tie and excluded points.

• theBy ties section is used to select points depending on number of point’s coordinatesmeasurements on image:

○ single – points not measured on any stereopair;

○ stereopair – points measured on at least one stereopair;

○ triplet – points measured at least in one triplet;

○ strip tie – points measured in interstrip overlap;

○ non-transferred – tie points not measured on adjacent images;

○ Used – tie points used in adjustment;

○ Excluded – tie points excluded from adjustment;

○ All – allows to select/unselect all points mentioned in the section.

• the Auxiliary section is used to select points depending on their type:

○ Geodetic – tie points which are not targeted and projection centers;

○ Targeted – targeted points;

○ Projection centers – projection centers;

○ with changed type – points with type that was changed during adjustment.

• the By residual value section allows to select points depending on residual valueon the point:

○ Worst points – points with maximal residuals (number of points to be displayedis specified in the appropriate field);

○ Over acceptable residuals – points with residuals, exceeding threshold.

Acceptable residuals are setup in the Parameters window on the Report tab (see Sec-tion 7.4 or Section 8.4).

Points selection by residuals value is performed using additional checkboxesGround XY, Ground Z, Tie in stereopairs XY, Tie in stereopairs Z, Tie on im-ages.

Residuals are also could be displayed using checkboxes in the Show errors section ofthe Display options window.

13


• In selectedmodels – used to select points located in selection area of stereopair/stripor multiple steropairs/strips:

○ At least one – points with measurements located in one selection area;

○ All – points with measurements located in all selection areas.

To Apply the specified settings of point display, click the appropriate button withoutclosing the Point filter window.

Ensure that the button is enabled in the Display options window.

5.4. Points selection

Triangulation points on a block scheme could be selected using standard group selectiontools. To change the selection mode are used the Edit › Group selection menu andthe Tools toolbar. (see the Objects selection chapter of the “Vectorization” UserManual).

Also the system provides possibility to highlight points of selected types on a blockscheme.

Click the button to select points to be highlighted. The Select points window isopened.

Fig. 4. The “Select points” window

To display points on a block scheme special – symbols.

The Select points window contains the following sections which are used to makepoints selection using different criteria:

14


vectorization.pdf

• the By coordinates section is used to select points depending on their coordinates:

○ ground control points (ground and projection centers);

○ check points (ground and projection centers);

○ excluded – points with known coordinates, excluded from adjustment;

○ no coordinates – points without geodetic coordinates (if the checkbox is set off,only ground control points are displayed);

○ all – allows to set on/off all checkboxes of section.

XYZ,XY and Z points checkboxes allows to choote type of displayed ground control,tie and excluded points.

• theBy ties section is used to select points depending on number of point’s coordinatesmeasurements on image:

○ single – points not measured on any stereopair;

○ stereopair – points measured on at least one stereopair;

○ triplet – points measured at least in one triplet;

○ strip tie – points measured in interstrip overlap;

○ non-transferred – tie points not measured on adjacent images;

○ Used – tie points used in adjustment;

○ Excluded – tie points excluded from adjustment;

○ All – allows to select/unselect all points mentioned in the section.

• the Auxiliary section is used to select points depending on their type:

○ Geodetic – tie points which are not targeted and projection centers;

○ Targeted – targeted points;

○ Projection centers – projection centers;

○ with changed type – points with type that was changed during adjustment.

• the By residual value section allows to select points depending on residual valueon the point:

15


○ Worst points – points with maximal residuals (number of points to be displayedis specified in the appropriate field);

○ Over acceptable residuals – points with residuals, exceeding threshold.

Acceptable residuals are setup in the Parameters window on the Report tab (see Sec-tion 7.4 or Section 8.4).

Points selection by residuals value is performed using additional checkboxesGround XY, Ground Z, Tie in stereopairs XY, Tie in stereopairs Z, Tie on im-ages.

Residuals are also could be displayed using checkboxes in the Show errors section ofthe Display options window.

• In selectedmodels – used to select points located in selection area of stereopair/stripor multiple steropairs/strips:

○ At least one – points with measurements located in one selection area;

○ All – points with measurements located in all selection areas.

The Select, Add, Subtract and Invert buttons are used to perform correspondingoperations with selected types of points.

5.5. The “Point attributes” window

The system provides possibility to view such attributes of triangulation points as type,coordinates and residuals. The Point attributes window is used to do this.

To open the Point attributes window click the button on the Adjustment toolbar.

16


Fig. 5. The “Point attributes” window for selected point

The Point attributes window consists of two sections: in the upper part of the windowthe editing panel is displayed; in the lower part of the window there is informationalpanel.

The Selected point field shows a name of highlighted point if a single point is highlighted.Or it contains a message about number of highlighted points.

A type of highlighted point is checked in the appropriate field of the Type by coordinates section.

The Type by coordinates section contains the following buttons:

• – allows to set selected point as a GCP;

• – allows to set selected point as a check point;

• – allows to exclude selected check point or GCP from further adjustment;

• – allows to add excluded tie point to adjustment;

• – allows to exclude tie point from further adjustment.

When a group of points of the same type (check, for instance) is selected, buttons allow tochange type of all selected points.

17


In order to view the selected points in the Triangulation points window, click theCoordinates button in the Type by coordinates section. The Triangulation pointswindow opens.

In the GCP list and Tie points list tabs of the Triangulation points window, only se-lected points will be shown.

If ground control points were among the selected points, the GCP list tab will be initially openin the Triangulation points window. Otherwise, the Tie points list tab will be first open in theTriangulation points window.

Double-click on point name in the Triangulation points window allows to open the Pointsmeasurement module to edit point’s position.

In order to change the coordinates of the selected points, click the Measure buttonin the Type by ties section. The Points measurement and the Triangulation pointswindows are opened (see the “Aerial triangulation” User Manual).

Only those selected points coordinates of which are to be changed will be shown in theGCP list and Tie points list tabs of the Triangulation points window.

If ground control points were among the selected points, the GCP list tab will be initially openin the Triangulation points window. Otherwise, the Tie points list tab will be first open in theTriangulation points window.

Section in the lower part of the Point attributes contains information about selectedpoint.

Information field contains any information only when just one point is selected on a block scheme.

In the section are displayed the following settings of selected point:

• Geodetic – information about GCP and check points:

○ reference coordinates – initial (user defined) coordinates of point;

○ reference std. dev. – initial weights of selected point coordinates;

○ adjusted coordinates – coordinates of point, obtained after adjustment;

○ residuals.

• Image residuals:

○ residuals (max) – maximal residuals on images for selected point;

○ number of images, where selected point coordinates were measured;

18


measurement.pdf

○ images – displays number of images where point’s coordinates were measured.For each of the images tie residuals are displayed (to average adjusted position)and also point’s measurements on image, coordinates of projection on image ofpoint’s adjusted position. For GCP and check points there are also coordinates ofprojection of initial geodetic coordinates on image.

• Stereopair residuals:

○ residuals (max) – maximal residuals on images for selected point;

○ number of stereopairs, where selected point coordinates were measured;

○ by stereopairs – displays residuals from mean value and mutual residuals foreach stereopair with selected point.

• Type – information about point’s type, whether it is a projection center, a code ofselected point (image code for projection centers) is also shown here);

• Images – list of images, where selected point coordinates were calculated;

• Strips – list of strips, where selected point coordinates were measured (on one imageat least).

5.6. List of triangulation points

The system provides possibility to view, search and edit of triangulation points. TheTriangulation points window is used to do this.

To open the Triangulation points window click the button on the Adjustmenttoolbar or choose Window › Triangulation points coordinates .

Fig. 6. The “Triangulation points” window

19


The Triangulation points window contains the toolbar with buttons used to performthe following operations:

• – allows to set on/off the filter of triangulation points display;

• – allows to setup the filter of triangulation points display;

• – allows to display all/selected on a block scheme points in the table;

• – is used to search for an image by name (part of name);

• – allows to display GC/check points initial coordinates (X, Y, Z – see the ’Aerialtriangulation” User Manual);

• – allows to display triangulation points coordinates obtained after adjustment (X',Y', Z');

• – allows to display errors on points:

○ dX, dY, dZ – error values on point;

○ Ex(adj.), Ey(adj.), Exy(adj.) – average residuals of points on images;

○ Esp_x(adj.), Esp_y(adj.), Esp_z(adj.) – tie residuals between stereopairs.

• – is used export the table content to text file with *.csv extension;

• – allows specify the following Error settings:

Fig. 7. The “Error settings” window

○ to set the following units of Ex(adj.), Ey(adj.) and Exy(adj.) errors displaying –for the central projection projects:

■ pixels;

■ millimeters.

For satellite scanner imaging and ADS 40/80/100 projects, error data are displayed onlyin pixels.

20


measurement.pdf

measurement.pdf

○ to select method of calculation tie residuals calculation between stereopairs(Esp_x(adj.), Esp_y(adj.), Esp_z(adj.)):

■ between models – tie residuals from difference of the same point positions,calculated on each stereopair;

■ from mean value – tie residuals from difference of the same point position,calculated on each stereopair and its mean adjusted position.

TheRefresh button is used for refreshing data in triangulation points table after changes.

The list of triangulation points is a table with the following columns:

• Code – an object code;

• Name – image name;

• Type – displays point’s type (tie, GCP, check, excluded);

• X, Y, Z – GC/tie point initial coordinates ( );

• X', Y', Z' – adjusted coordinates of triangulation points ( );

• dX, dY, dZ – error values on point ( );

• Ex(adj.), Ey(adj.), Exy(adj.) – average residuals of points on images ( );

• Esp_x(adj.), Esp_y(adj.), Esp_z(adj.) – tie residuals between stereopairs ( ).

The Triangulation pointswindow is synchronized with 2D-window: when point selectingin 2D-window it is also selects in the table. Double-click by row with point allows to openthe Points measurements module to edit point’s position.

The system provides possibility to edit position of triangulation points selected in 2D-window in the Points measurementsmodule. Perform the following actions to do this:

1. Select points in 2D-window with one of the group selection instruments (see the“Vectorization” User Manual).

2. In the Triangulation points click the button to show in the table only points se-lected in 2D-window.

3. In the Point attributes window click the Measure button. The Points measure-ment module opens (see Section 11).

In the Triangulation points window is displayed list of points, selected in 2D-win-dow.

21


vectorization.pdf

4. Edit points position and click OK to return to the adjustment step.

6. Brief residuals reportThe system allows to display brief residuals report, which contains RMS errors, meanabsolute and maximal adjustment errors.

Brief residuals report is used for quick estimation of adjustment results without viewingdetailed report.

In order to display brief residuals report, click the button on the Adjustment toolbar.The Residuals window opens.

Fig. 8. The “Residuals” window

The brief report contains information about mean absolute errors on GCP and checkpoints (in meters), about tie and targeted points, as well as about tie residuals on pro-jection centers (mutual and from mean residuals).

If adjustment is performed by bundle adjustment procedure, the Sigma_0 value is calculated.

The Sigma_0 value shows howmuch adjustment discrepancies correspond to specified thresholdson points coordinates measurements and input GCP data (GCP and projection centers coordin-ates). If the thresholds were specified correctly, the Sigma_0 has value close to 1 (±30%).

If the Sigma_0 is much greater than 1, there are errors in points coordinates measurements, ininput GCP data or because of incorrectly specified thresholds on points coordinates measure-ments. If the Sigma_0 is mush less than 1, the threshold on points coordinates measurementswere specified incorrectly or there are some other errors (see Section 10).

22


7. Adjustment of central projection images block

7.1. Workflow during adjustment of central projection imagesblock

The system allows to process images of central projection.

Central projection images include airborne images, close range images, satellite imagesof central projection (see “Project creation” User Manual).

Setup adjustment parameters of central projection images block, selection of adjustmentmethod and adjustment model is performed using the Parameters window.

To open the Parameters window click the button on the Adjustment toolbar.

Fig. 9. Parameters of adjustment of central projection images block

In order to setup adjustment parameters perform the following actions:

1. Specify coordinate system of a project on the Coordinate system tab.

2. Select the adjustment method and setup main parameters of adjustment on theAdjustment tab.

3. Setup display of adjustment results on report and define adjustment residualsthreshold using the Report tab.

4. Click the Compute button to start the adjustment operation or click OK to save thesettings and close the Parameters window.

23


project.pdf

7.2. The “Coordinate system” tab

The Coordinate system tab allows to specify coordinate system, if it wasn’t specifiedon project creation stage, or allows to change current coordinate system (see “Coordinatesystem selection” chapter in the “Project creation” User Manual).

It is not necessary to define coordinate system for block adjustment in free model.

Fig. 10. Parameters of adjustment

7.3. The “Adjustment” tab

7.3.1. Adjustment parameters of central projection images block

The Adjustment tab allows to choose method and model of adjustment of central pro-jection images block, and also to setup adjustment parameters.

24


project.pdf


In order to setup adjustment parameters of central projection images check if georefer-ence data is available (measured GCP coordinates or exterior orientation parameters)and perform the following actions:

1. Click the button on the Adjustment toolbar. The Parameters window opens.

2. Click the Adjustment tab.

3. In case of georeference data is unavailable or not enough, it is recommended toperform adjustment in free model. To do this set the Free model checkbox andselect one of the following georeference models:

• free – during adjustment the system performs rotation, zoom and shift of projectioncenters coordinates in such a way that X,Y,Z coordinates of projection center ofthe first image of block are 0,0,0, of the second – b,0,0, where b – value of basisin meters specified in the Basis field;

• by block scheme – during adjustment the system performs rotation, zoom andshift of projection centers coordinates in accordance with block layout.

Adjustment in free model allows to perform control of blunders in tie points measurements.

4. In the Approximation computing method section choose method and specifyparameters of first approximation:

• independent strips – during adjustment independent models by images in eachstrip are creating (using tie points between images) and further joining to common

25


model. After that performs orientation by GCP in selected geodetic coordinatesystem;

It is necessary to have enough amount of points measured in triplets to use method ofindependent strips. This allows to quickly adjust block which contains big number ofimages.

• by block scheme – adjustment using current block scheme of block layout forapproximate adjustment considering results of all block measurements.

5. [optional] In order to use previous adjustment results in new adjustment session,set the Use current solution checkbox. This allows to use results obtained duringprevious step of adjustment and not calculate the values again.

It is used in those cases when iterations of adjustment disagree, some measurements arecorrected and the process starts again, for example, when there are lack of points meas-urements in triplets for images blocks acquired from UAV or they are insufficient.

6. In the Adjustment method section choose method and specify parameters of se-lected method:

• keep initial approximation – aerial triangulation using initial approximationresults;

The method is used to search for blunders, such as incorrectly specified coordinates ofGCP.

For “rough” adjustment of images block acquired from UAV, set the keep initial approx-imation checkbox and adjust block using the Initial approximation methodby blockscheme.

• independent stereopairs – preliminary aerial triangulation by stereopairs of in-dependent models and their further merging into common block and orientationof the block network in geodetic coordinate system;

• bundle adjustment – aerial triangulation and block adjustment by all images atonce is used for final adjustment.

7. In the Tie point rejection section specify conditions of triangulation points filtering:

• worst points – to filter specified number of tie points with maximal residuals;

• over acceptable residuals – to filter tie points where residuals values exceedpreset threshold (on the Report tab).

26


8. In the Reject on one iteration input field specify number of points to be rejected(in accordance with specified filtering conditions) on each step of iterative procedureof adjustment.

9. [optional] For automatic calculation and compensation of systematic measurementerrors click the Systematic errors compensation button.

10. [optional] For automatic calculation of camera parameters during adjustment (if inproject using incomplete set of camera data or when it is necessary to refine thedata), set the Perform self-calibration checkbox and click the Self-calibrationparameters button.

11. [optional] In order to select project images for adjustment set the Adjust subblockcheckbox and click the Select subblock button.

12. [optional] TheConsistency check checkbox allows to check availability of all sourcedata that are required to adjustment. If any data is not available, the error messagedisplays and adjustment process doesn’t start.

If the Approximation computing method by a block scheme and the bundle adjustmentmethod were chosen, consistency check doesn’t perform regardless of the checkbox.

13. Click OK to save parameters or Compute to start the adjustment process.

7.3.2. Adjustment in free model

Adjustment in free model allows to adjust a block without georeference data. Adjustmentin free model to perform control of blunders in tie points measurements.

To perform block adjustment in free model perform the following actions:



3. Set the Free model checkbox.

4. Select one of the following types of model georeference:

• free – during adjustment the system performs rotation, zoom and shift of projectioncenters coordinates in such a way that X,Y,Z coordinates of projection center ofthe first image of block are 0,0,0, of the second – b,0,0, where b – value of basisin meters specified in the Basis field;

• by block scheme – during adjustment the system performs rotation, zoom andshift of projection centers coordinates in accordance with block layout.

27


5. Specify the rest of adjustment parameters.

After interior and relative orientation it is necessary to measure GCP coordinates andadjust a block in selected coordinate system.

In order to adjust a block in selected coordinate system after its adjustment in freemodel, perform the following actions:

1. Adjust a block in free model.

2. Measure coordinates at least of three GCP or load data about at least of threeprojection centers.

3. On the Adjustment tab of the Parameters window set the Free model checkboxoff.

4. Set the Use current solution checkbox.

5. In the Adjustment method section select the Keep initial approximation option.

6. Adjust the block. All data of project processing (vector objects, TIN and DEM) createdin free model, transform to selected geodetic coordinate system.

7.3.3. Independent strips method

Independent strips method is used to perform first approximation. This method consistsin preliminary creation of independent models using images of each strip (by tie pointsmeasured between images) and further joining of the models into integrated model.Then the model is orientated by GCP in geodetic coordinate system.

Independent strips method is applied in general to reveal blunders, such as incorrectlyspecified GCP coordinates or errors in positioning of tie points.

Prior to adjustment using independent strips method check a direction of coordinatesystem. If the coordinate system is turned at 180° in relation to geodetic coordinatesystem, perform one of the following actions:

• reverse the strips order;

• reverse images in the strip order;

• turn all images to 180°;

• use another coordinate system with reversed axis orientation;

• change signs of X and Y coordinates of all GCP to opposite (only when perform theadjustment in Cartesian coordinate system).

28


In order to calculate initial approximation using independent strips method perform thefollowing actions:



3. In the Approximation computing method section select the independent stripsmethod.

4. Click the Initial approximation parameters button. The Independent strips ad-justment parameters window opens.

Fig. 12. The “Independent strips adjustment parameters” window

5. Specify the Maximum points distance (in bases) parameter. The parameter isused to process close range projects, where points with big X-parallax differenceis measured.

Applying this parameter results in automatic excluding from adjustment on each stereopairthose points, distance from which to projection centers exceeds specified distance in bases.For example, iа a stereopair basis is 10 meters, and predefined value of maximal distanceis 100meters, then the systemwill exclude the points which distance from projection centersis more than 1000 meters.

It is recommended to set the value of the parameter not more than 100 bases, since whenpoints with distance exceeding 100 bases of survey are considered in adjustment, inde-pendent strips (or stereopairs) method may disagree or bring to result with big errors.

If there are no blunders in measurements, decreasing of the Maximum points distance(in bases) value results in decreasing of RMS on a block. Too small value results inbreaking of adjustment operation and shows warning about error due to lack of tie pointsin triplet.

29


6. In theMaximum strip length (number of images) input field set number of images,on which each strip will be divided. This parameter is used to decrease a numberof systematic errors inside strips, which bring to adjustment blunders when stripsare more than 20-30 images in length.

It is recommended to set the value of the parameter within 10-30 images. Strips that havemore images will be divided into part of specified size during adjustment procedure.

7. Set theAdjustment accuracy slider to a value, at which the system should completeiterative adjustment procedure.

It is recommended to use the value 0.5 (it is set by default). When it equals 0.5-0.55, theadjustment process converges and provides an optimal ratio “accuracy-time”. If you increasethe value more than 0.5, the accuracy generally does not improve considerably, but pro-cessing time may increase very much up to infinite looping of iterations (in this case youmay use the Stop button to stop the adjustment process). If the accuracy value is less then0.5 the adjustment time is reducing due to accuracy decreasing. The algorithm of optimiz-ation has a complex nature, that is why the dependence of adjustment time and accuracyfrom this parameter value is nonlinear.

8. [optional] In the Equation weights field input the values of weights for GCP andtie points, as well as for projection centers to reduce adjustment errors.

Considerable change of equation weights values leads to reducing of actual adjustmentaccuracy. There is some optimal value, at which the accuracy is maximal. In most casesthe weight value of 1 is near to optimal. It is necessary to have check points enough tocontrol the result of optimal value search. If there is a lack of check points it is recommendedto temporarily transfer some GCP to check points (see Section 11).

In order to record numbers of high/low order an exponent is used, for example: “5.7e-4”.

9. [optional] In order to useGCPmeasurements during adjustment, set theUse groundcontrol point coordinates checkbox on.

10. [optional] In order to use projection centers data during adjustment, set the Useprojection center coordinates checkbox on.

7.3.4. Method of initial approximation by a block scheme

Method of initial approximation calculation by a block scheme consists in using of blocklayout for block adjustment (see the “Aerial triangulation” User Manual).

Initial approximation by block scheme with adjustment by bundle adjustment method is usedfor a single image adjustment.

To calculate initial approximation by block scheme perform the following actions:

30


measurement.pdf



3. In the Approximation computing method section select the by block schememethod.

4. Click the Initial approximation parameters button. The Initial approximationparameters window opens.

Fig. 13. The “Initial approximation parameters” window

5. Setup initial approximation parameters by block scheme:

• Use ground control point coordinates – allows to use GCPmeasurements foradjustment operation;

• Use projection center coordinates – if projection centers coordinates areavailable in catalogue of exterior orientation parameters it is possible to use themfor adjustment operation;

• Use exterior orientation angles – if exterior orientation angles are available incatalogue of exterior orientation parameters, the option allows to use them foradjustment operation;

• Perform direct georeferencing – allows to calculate tie points coordinates usingexterior orientation data. If the checkbox is set off, then X,Y coordinates of tiepoints are calculated using frames of images on a block scheme. For Z coordinatethe system assigns average terrain height from heights difference data, specifiedin a project properties (see the “Project creation” User Manual).

6. Click OK. All the missing data is calculated by frames of images of current blockscheme, considering average terrain height from heights difference data, specifiedin a project properties (see the “Project creation” User Manual).

31


project.pdf

project.pdf

7.3.5. Independent stereopairs method

Independent stereopairs method consists in preliminary creation of single independentmodels by stereopairs, and further joining them into common block and orientation ofthe block in geodetic coordinate system using GCP.

Independent stereopairs method is used to enhance accuracy of adjustment results, obtainedduring calculation of initial approximation, in order to detect smaller errors and, in case of satis-fying results, for final adjustment.

For adjustment using the independent stereopairs method perform the following actions:



3. In the Adjustment method section select the independent stereopairs method.

4. Click the Adjustment parameters button. The Independent stereopairs adjust-ment parameters window opens.

Fig. 14. The “Independent stereopairs adjustment parameters” window

5. Specify the Maximum points distance (in bases) parameter. The parameter isused to process close range projects, where points with big X-parallax differenceis measured.

Applying this parameter results in automatic excluding from adjustment on each stereopairthose points, distance from which to projection centers exceeds specified distance in bases.For example, iа a stereopair basis is 10 meters, and predefined value of maximal distanceis 100meters, then the systemwill exclude the points which distance from projection centersis more than 1000 meters.

32


It is recommended to set the value of the parameter not more than 100 bases, since whenpoints with distance exceeding 100 bases of survey are considered in adjustment, inde-pendent strips (or stereopairs) method may disagree or bring to result with big errors.

If there are no blunders in measurements, decreasing of the Maximum points distance(in bases) value results in decreasing of RMS on a block. Too small value results inbreaking of adjustment operation and shows warning about error due to lack of tie pointsin triplet.

6. In the Optimization algorithm section choose one of the following algorithms:

• gradient (by default);

The adjust by subblocks checkbox allows to divide the image block into subblocksduring the adjustment in order to increase the system performance (that in certain situ-ations may affect the adjustment accuracy). It is recommended to use this function whenadjusting blocks of areal survey with a large number of images and about 80% overlap.

• Matrix inversion.

Method of matrix inversion was used in previous versions of the system and it is de-veloped for relatively small blocks of images (up to 500 images).

7. Set theAdjustment accuracy slider to a value, at which the system should completeiterative adjustment procedure.

It is recommended to use the value 0.5 (it is set by default). When it equals 0.5-0.55, theadjustment process converges and provides an optimal ratio “accuracy-time”. If you increasethe value more than 0.5, the accuracy generally does not improve considerably, but pro-cessing time may increase very much up to infinite looping of iterations (in this case youmay use the Stop button to stop the adjustment process). If the accuracy value is less then0.5 the adjustment time is reducing due to accuracy decreasing. The algorithm of optimiz-ation has a complex nature, that is why the dependence of adjustment time and accuracyfrom this parameter value is nonlinear.

8. [optional] In the Equation weights field input the values of weights for GCP andtie points, as well as for projection centers to reduce adjustment errors.

Considerable change of equation weights values leads to reducing of actual adjustmentaccuracy. There is some optimal value, at which the accuracy is maximal. In most casesthe weight value of 1 is near to optimal. It is necessary to have check points enough tocontrol the result of optimal value search. If there is a lack of check points it is recommendedto temporarily transfer some GCP to check points (see Section 11).

In order to record numbers of high/low order an exponent is used, for example: “5.7e-4”.

33


9. [optional] In order to useGCPmeasurements during adjustment, set theUse groundcontrol point coordinates checkbox on.

10. [optional] In order to use projection centers data during adjustment, set the Useprojection center coordinates checkbox on.

7.3.6. Bundle adjustment method

Bundle adjustment method is used for final adjustment. When using the bundle adjust-ment method aerial triangulation and adjustment are performed for all images at once.

For adjustment by bundle adjustment method it is not necessary to have images triple overlapavailable.

The system provides possibility to perform adjustment of a single image by bundle ad-justment method with initial approximation by a block scheme. To do this it is necessaryto have at least three GCP available in the image. Besides, it is necessary to set theUse exterior orientation angles checkbox off in the Bundle adjustment parametersand Initial approximation parameterswindows, and set the Perform direct georefer-encing checkbox off in the Initial approximation parameters window.

For adjustment of block images using the Bundle adjustment method perform the fol-lowing actions:



3. In the Adjustment method section select the bundle adjustment method.

4. Click the Adjustment parameters button. The Bundle adjustment parameterswindow opens.

Fig. 15. The “Bundle adjustment parameters” window

34


5. In the Point on image measurement precision (RMS) section specify accuracyfor manual measurements and automatic measurements of points coordinateson images in pix or mm.

Accuracy of points coordinates measurements depends on photographic material quality(“grain”, blur, image deformations, not considered in camera parameters, including geometricdistortions of scanner), terrain type and human factor (accuracy of measurements performedby operator).

If step of image scanning corresponds to its quality and points are correctly recognized,measurements accuracy value lies within 0.5–1.0 of a pixel size.

6. In the Weights section define weights for the following data:

• weights of image measurements;

• ground control points – weight of GCP coordinates;

Input value of GCPweight is equivalent to multiplying of accuracy value of GCP coordin-ates taken from Catalogue of ground control points by 1/ W, where W – is a weightvalue. If GCP weight value is more than 1, then binding of triangulation network to GCPcoordinates is strong, if the weight values are less than 1, the binding is weak (see the“Aerial triangulation” User Manual).

• projection centers – input field for weight of projection centers coordinates.Accuracy of projection centers coordinates are setup in the Catalogue of exter-ior orientation parameters (see the “Aerial triangulation” User Manual);

• exterior orient. ang. (omega, phi, kappa) – input field for weight of exteriororientation angles. Accuracy of exterior orientation angles are setup in theCatalogue of exterior orientation parameters (see the “Aerial triangulation”User Manual).

7. [optional] In order to use GCP measurements for adjustment by default, the Useground control point coordinates checkbox is set on. Set the checkbox off andthe system will not use the data for adjustment.

8. [optional] In order to use projection centers coordinates for adjustment by default,the Use projection center coordinates checkbox is set on. Set the checkbox offand the system will not use the data for adjustment.

9. [optional] In order to use exterior orientation angles for adjustment by default, theUse exterior orientation angles checkbox is set on. Set the checkbox off and thesystem will not use the data for adjustment.

10. Click OK to complete parameters setup.

35


measurement.pdf

measurement.pdf

measurement.pdf

7.3.7. Bundle adjustment with self-calibration

Method of bundle adjustment with self-calibration is used if a project contains not fullset of camera parameters (for example, there are no all necessary passport data ofcamera), or if it is necessary to clarify camera parameters (see the “Aerial triangulation”User Manual).

Self-calibration of camera parameters – is automatic calculation of camera parameters (correc-tions to principal point coordinates and focal length, distortion coefficients) during adjustmentoperation.

Self-calibration of camera parameters is performed only during the bundle adjustment. Whenexecute adjustment using other methods, specified corrections to camera parameters are con-sidered but not optimized.

If interior orientation is done incorrectly due to absence of some camera data or due toincorrectly specified camera data, so self-calibration of camera parameters on adjustmentstep using georeference data (GCP, exterior orientation parameters) allows to calculatemissing data.

For self-calibration of camera parameters at least focal length is required, specified in theCamera window (see the Interior orientation chapter in the “Aerial triangulation” User Manual)

For self-calibration of camera parameters it is recommended to use georeference data (GCPand/or exterior orientation parameters) during block adjustment. The system also allows to useself-calibration in free model, that allows to analyze systematic errors and make decisionwhether it is necessary to go on with self-calibration.

Self-calibration results depend on a set of initial camera data, block configuration (position ofblock images and strips, which depends on flight scheme), availability of georeference data(GCP, exterior orientation parameters) and tie points measurements. Description of all self-calibration features are out of scope of the User Manual (see appropriate references about thesubject).

To use self-calibration of camera parameters during adjustment operation perform thefollowing actions:



3. Set the Perform self-calibration checkbox on in the Camera parameters self-calibration section.

4. Click the Self-calibration parameters button. The Camera parameters self-cal-ibration window opens.

36


measurement.pdf

measurement.pdf

Fig. 16. The “Camera parameters self-calibration” window

5. In the Cameras table mark cameras which are needed to be calibrated.

6. Select the Calibration type: Physical or Mixed.

For self-calibration of a camera having a focal plane shutter, it is recommended to useMixed self-calibration.

7. Click the button to copy coefficients from the camera passport.

8. In the Coefficients table define initial values of camera parameters in the Initialvalue column and select parameters for optimization in the Optimize column (seeSection 7.3.7).

9. Click OK to complete setup of camera self-calibration parameters.

10. Setup the adjustment parameters in the Parameters window.

11. Click the Compute button.When the adjustment procedure is completed estimatethe adjustment accuracy and results of camera self-calibration shown in theAdjust-ment column of the Camera parameters self-calibration window. In case of sat-isfying results pass to the next step, otherwise, change initial values and/or a setof optimization parameters and adjust the block once more.

12. Click the Save button to save the adjustment results.

13. Choose Orientation › Manage cameras (Ctrl+Alt+I). The Manage project cam-eras window opens.

37


14. In the Project cameras list select new calibrated camera (cameras) [selfcal].x-cam and assign it (them) to project images. Click the Execute button (see the“Aerial triangulation” User Manual). Click OK.

15. Choose Orientation › Interior orientation › Calculate interior orientation.

16. Click the button Settings on the Adjustment toolbar. Block will be adjustedconsidering new results of interior orientation.

7.3.8. Parameters of camera self-calibration

The system allows to setup parameters of camera self-calibration. The Camera para-meters self-calibration window is used to do this. The window also allows to specifyinitial values of parameters and define a set of parameters for optimization during ad-justment operation.

Fig. 17. The “Camera parameters self-calibration” window

The Camera parameters self-calibration window contains the following tables:

• the Cameras is used to select cameras to be calibrated;

• the Coefficients is used to define initial parameters values, to select optimizationparameters and perform analysis of parameters values calculated during adjustment;

• a toolbar that contains buttons used for work with data of the Coefficients table.

The Calibration type section allows to choose method of camera self-calibrationcoefficients calculation. In most cases it is recommended to choose the Physical calib-ration type.

38


measurement.pdf

For self-calibration of a camera having a focal plane shutter, it is recommended to use Mixedself-calibration.

The Coefficients table contains the following columns:

• Parameter – list of the following camera parameters:

○ Dx0 – offset from the principal point by X in mm;

○ Dy0 – offset from the principal point by Y in mm;

○ DF – deviation from focal length value in mm;

○ K1, K2, K3 – coefficients of radial distortion;

○ P1, P2 – coefficients of tangential distortion;

○ b1, b2 – coefficients of image deformation.

• Initial value – initial parameters values, that are used during self-calibration in adjust-ment operation;

• Optimize – selection of parameters for optimization during adjustment operationconsidering initial values.To enable parameter optimization, click on the , after thatthe “+” sign is shown in the column. In order to unselect optimization parameter it isnecessary to use one more mouse click;

• Camera – initial data of camera, specified on interior orientation step in the Camerawindow (see Orientation › Manage cameras);

On step of initial orientation during input of camera parameters, user specifies absolute valuesof principal point coordinates and focal length, that is why data for Dx0, Dy0, DF deviations areabsent in the Camera column.

If table distortion data was setup during step of camera parameters input, so in the Cameracolumn distortion coefficients are shown for this camera. The coefficients are obtained as aresult of approximation of initial table distortion data using the following formula:

xcorr = x − k1xr2 − k2xr

4 − k3xr6 − P(2x2 + r2) − 2P2xy − b1x − b2y;

ycorr = y − k1yr2 − k2yr

4 − k3yr6 − P(2y2 + r2) − 2P2xy − a1x − a2y;

r2 = x2 = y2,

where:

x, y – coordinates of symmetry point (zero distortion) in mm;

k1, k2, k3 – coefficients of radial distortion;

p1, p2 – coefficients of tangential distortion;

39


a1, a2, b1, b2 – image deformation coefficients.

As a result of approximation by formula that uses least-squares method, the distortion coef-ficients and root-mean-square error (RMS) of approximation are calculated.

• Adjustment – results of self-calibration (after adjustment operation), calculated asa result of adjustment of selected for optimization parameters.

The Camera parameters self-calibration window contains the following buttons usedto work with the data of the Coefficients table:

• – allows to select all parameters for optimization;

• – allows to cancel optimization of all parameters;

• – allows to setup “0” value for all parameters in the Initial value column;

• – allows to copy all parameters values from the Cameras column to the Initialvalue column;

• – allows to copy all parameters values from the Adjustment column to the Initialvalue column;

• – allows to copy parameter value from selected cell to the Initial value column.

By default during the very first self-calibration of project cameras is used the followingscenario:

• all project cameras are selected in the Cameras table for self-calibration;

• initial camera parameters are used as initial values of each camera parameters forself-calibration (i.e. The data from the Cameras column to the Initial value columnare copied automatically);

Since absolute values of principal point coordinates and focal length are specified at cameraparameters input, then in the Initial value column for Dx0, Dy0, DF deviations the “0” value isshown.

• the following parameters are selected for optimization during adjustment procedure:DF to calculate focal length deviation from its initial value, and K1, K2, K3 to calculatedistortion coefficient.

40


Fig. 18. Initial table data of distortion from camera passport, defined in the “Camera” window during inputof camera parameters step

41


Fig. 19. Distortion coefficients K1, K2, K3 calculated as a result of approximation by formula, and distortionerror

7.3.9. Compensation of systematic errors

There is a possibility to calculate systematic errors of block measurements automaticallyand compensate them during block adjustment.

42


Fig. 20. The “Systematic error compensation” window

The Projection center coordinates (GPS) section is used to specify Compensationtype option, to define a type of compensation to be applied to coordinates of projectioncenters (GPS):

• do not apply;

• common for block – is used to consider systematic errors in each block;

• common for strip – is used to consider errors obtained due to work of GPS device(recommended option).

It is possible to include calculated values of GPS systematic error to adjustment report.

43


If a survey was performed using GPS equipment, that was used to define projectioncenters coordinates, it is possible to calculate corrections to these coordinates duringadjustment. To do this set the appropriate checkbox in the Projection center coordin-ates (GPS) section. In general the system considers errors, added by GPS deviceduring flight along a strip, that is why it is recommended to choose the common forstrip option.

The XY and Z checkboxes allow to specify coordinates to which corrections will be ad-ded.

The Exterior orientation angles (constant rotation) section is used to specify a typeof corrections applied to exterior orientation angles of images, if they are specified in aproject and are considered for adjustment:

• do not apply;

• common for block – is used to consider systematic errors in each block;

• common for strip – is used to consider errors obtained due to work of GPS device(recommended option).

Corrections in the Exterior orientation angles (constant rotation) section are used only forbundle adjustment method.

The corrections are as follows:

Δ ω = AαΔ ϕ = Aϕ ,Δ κ = Aκ

where A – coefficients calculated at adjustment.

7.3.10. Select subblock

The system allows to perform adjustment just for a part of a block (subblock).

To perform subblock adjustment perform the following actions:



3. Set the Adjust subblock checkbox.

4. Click the Select subblock button. The Select subblock window opens.

44


Fig. 21. The “Select subblock” window

5. Select images to define a subblock content, using the following tools of imagesselection:

• – allows to select all images in the list;

• – allows to deselect all images in the list;

• – allows to inverts images selection;

• – allows to select highlighted images;

• – allows to unselect highlighted images;

• – allows to select the images that are highlighted on block scheme in 2Dwindow;

• – allows to highlight on block scheme in 2D window the images selected inthe list.

6. Click OK.

7.4. The “Report” tab

The Report tab is intended for setup of adjustment results display, as well as for setupparameters of errors display on a block scheme.

45


Fig. 22. The “Settings” window

The Report tab contains the following parameters:

1. The parameters group Include in the report is used to setup adjustment resultsdisplay in report:

46


• Residuals – in report display the following types of residuals :

○ control, check – adjustment residuals on GCP and check points;

○ on stereopairs – adjustment residuals on points, which coordinates arechanged at least on a single stereopair:

■ tie and pass residuals – tie residuals, calculated between strips and insidestrips;

■ include singles – adjustment residuals on points, which coordinates aremeasured on a single stereopair.

It is recommended to set the include singles checkbox on only when it is necessaryto calculate the from mean value residuals in the Tie residuals section.

During adjustment by independent stereopairs method or by independent stripsmethod residuals on points measured just on a single stereopair equals to zero.

During bundle adjustment residuals on points measured just on a single stereopairwere less then 0.1 of vertical parallax value on terrain.

○ on images – adjustment residuals on images (in mm or pixel);

○ exterior orientation angles – residuals on angle exterior orientation paramet-ers, only if the angles are specified in a project were used during adjustment;

○ details – is used to display detailed information about residuals (including im-ages and stereopairs numbers);

○ mark bad points – is used to display adjustment residuals, which exceedspecified threshold;

○ include acceptable – is used to display all adjustment residuals, includingthose that fall within specified threshold.

• Exterior orientation parameters – is used to display exterior orientation para-meters of each image in a report;

• Catalog – is used to display in a report geodetic coordinates of all points calcu-lated during adjustment. When the print residuals checkbox is set, the pointscatalogue shows residuals on GCP and check points;

• GPS coordinates correction – is used to display GPS corrections to projectioncenters in a report;

It is recommended to set the GPS coordinates correction checkbox when projectioncenters are used as GCP, and systematic error in GCP coordinates is defined.

47


• Exterior orientation angles correction – is used to display corrections to angularexterior orientation parameters in a report;

It is recommended to set the Exterior orientation angles correction checkbox on whenexterior orientation angles are used during adjustment, and to define systematic errorsof exterior orientation angles.

• Self-calibration results – is used to display self-calibration results of cameraparameters in a report.

It is recommended to set the Self-calibration results checkbox on only for bundle ad-justment operation. In this case it is recommended to set the Perform self-calibrationcheckbox on in the Camera parameters self-calibration section.

2. The Tie residuals section is used to select method of calculation tie residualsbetween stereopairs:

• between models – tie residuals from difference of the same point positions,calculated on each stereopair;

• from mean value – tie residuals from difference of the same point position, cal-culated on each stereopair and its mean adjusted position.

3. The Acceptable residuals section allows to specify residuals threshold of blockadjustment:

It is necessary to specify acceptable residuals of block adjustment in units of project co-ordinate system.

• all the same – the same thresholds for all kinds of points. In the All section thereare the XY and Z fields used to input residual threshold;

• by point type – different threshold for each types of points. In the the Groundcontrol, Check, Projection centers, Tie, Targeted, Tie-projection centersfields there are the XY and Z fields used to input threshold for each type of resid-uals.

4. The Scale section allows to specify survey scale, used for residuals re-calculationfrom image scale to block scale and vice versa.

A value of survey scale is calculated automatically during adjustment, or is specifiedmanually.

Mean pixel size on terrain (ground sample distance) calculated after adjustment,is specified in the Mean pixel size (GSD) input field.

48


5. The Residuals print format section is used to specify format of numeric values ofresiduals in a report:

• fixed point – residuals display with accuracy up to 3 decimal points;

• floating point – residuals display with accuracy up to 3 significant digits.

6. The Model section is used to specify format of exterior orientation angles system(see Appendix A):

• Alpha, omega, kappa – are used in Russian coordinate systems;

• omega, phi, kappa – are used in international coordinate system (see Ap-pendix A).

7. The Angle units section is used to setup format of angles measurements units:

• radians;

• degrees;

• gons – plane angle measurement unit that equals to 1/100 of flat right anglemeasurement unit, and full angle is 400 gons.

8. The Image residuals units section is used to setup format of angles measurementsunits:

• pixels;

• mm.

8. Adjustment of scanner blocks

8.1. Workflow of scanner block adjustment

At that preliminary parameters setup is required. A set of the parameters depends onmethod of scanner data processing: rigorous, generic or RPC-method (detailed inform-ation about algorithms of space images processing see in the “Project creation” UserManual).

For setup of adjustment parameters the system provides dedicated mode for adjustmentof scanner block, acquired by ADS 40/80/100 digital sensor.

The Parameters window is used to setup parameters of adjustment accuracy, selectadjustment method procedure, as well as to setup a content of adjustment results report.

49


project.pdf

Fig. 23. The “Parameters” window toolbar

A toolbar in the upper part of the Parameterswindow is intended to save/load paramet-ers settings on all tabs of the window and contains the following buttons:

• – allows to load parameters settings saved previously to active profile resources;

• – allows to save all parameters settings to *.x-ini file located in the ProjOptionproject folder in active profile resources;

• – allows to load parameters settings from <project name>.options.x-solver-s. filelocated in a folder of Windows file system;

• – allows to save parameters settings to <project name>.options.x-solver-s filelocated in a folder of Windowsfile system;

• – allows to restore default parameters settings.

In order to adjust scanner images block perform the following actions:

50


1. Click the button of the Block Adjustment toolbar. The Parameters windowopens.

2. On the Points tab select coordinate system and setup parameters of adjustmentaccuracy and points use.

3. On the Images tab choose method and parameters of adjustment.

4. Click the Apply to all button to apply adjustment settings to all project images orclick the Apply button to apply all adjustment settings to a single image selectedin the Images table.

5. On the Report tab specify values of the following acceptable residuals:

• for control, check and tie points;

• acceptable residuals (RMS) of grouping of images, acquired from the same orbit(if there was any grouping).

Setup display parameters of adjustment results in a report.

6. Click OK. After that the system checks parameters settings. If their values are in-correct, the tab containing field with incorrect value becomes active, and the inputfield gets red. If parameters were specified correctly, the Parameterswill be closed.

7. Click the button to start the adjustment operation.

8. To display and analysis of adjustment results in the report window click thebutton of the Block Adjustment toolbar.

9. In case of satisfactory results of adjustment click the button of theBlock Adjust-ment toolbar to save adjustment data. Proceed to step of project processing (DEMcreation, vectorization, orthomosaic creation). Otherwise (in case of poor result),correct errors appeared on a step of data collection for aerial triangulation (see the“Aerial triangulation” User Manual) and/or change settings of adjustment parametersin the Parameters window and start the adjustment operation once more.

8.2. The “Points” tab

The Points tab of the Parameters window allows to choose project coordinate system,and to setup parameters of calculation accuracy of points coordinates.

51


measurement.pdf


The Points tab contains the following parameters:

• the Coordinate system section is used to display project coordinate system. TheSelect... button allows to change current coordinate system of a project (detailed in-formation about selection of coordinate system see in the “Project creation” UserManual).

• the Points coordinates accuracy section is used to specify accuracy of pixel (onimage) and geodetic (on ground) points coordinates, included to adjustment:

○ Image measurements RMS (pix.) – a priori accuracy of measurements of GCPand tie points on images (if the latter were used for adjustment);

○ Unity-weighted geodetic coordinates accuracy – a priori measurement errorsof GCP coordinates in plane (X,Y) and height (Z), which weight equals 1.

52


project.pdf

Unity-weighted accuracy of plain and height geodetic coordinates are setup in the sameunits as those used in project coordinate system.

• thePoints usage section is used to setup using/excluding tie points in/from adjustmentoperation (the checkbox is set by default).

Coordinates delta threshold (pix) – minimal difference between point’s position in pixel co-ordinates, calculated on adjacent iterations using exterior orientation parameters during ad-justment operation. When the parameter is reached, iterative operation is terminated.

It is recommended to setup this threshold value in range from 0.01 to 0.2 pixel.

• the DTM for adjustment section is used to consider DEM or specified constant elev-ation in adjustment:

○ Do not use – to perform adjustment without DEM;

○ Use DEM – allows to use DEM from file (out from active profile resources);

The system memorizes the last used DEM .

○ Use constant elevation – allows to use constant elevation of relief instead of DEM.

It is recommended to perform adjustment of projects with images of flat terrain.

• the Convergence angle point rejection section allows to exclude from adjustmenttie points with small convergence angles. To do this set the Reject tie point withsmall convergence angle checkbox and specify theMinimum acceptable conver-gence angle in degrees.

During the adjustment, the points having convergence angles lesser thanMinimum accept-able convergence angle will be processed as tie points excluded from adjustment, but typeof these points will not be changed to Excluded (see the “Points filter” chapter of the “Aerialtriangulation” User Manual).

Set the Include the list of points rejected by convergence angle checkbox in the Reporttab to add the list of tie points excluded by convergence angle into the report

53


measurement.pdf

measurement.pdf

8.3. The “Images” tab

8.3.1. Parameters of scanner images adjustment

The Images tab is used to specify adjustment method of scanner images block, as wellas to setup the adjustment parameters.


The Images tab consists of the following sections:

• in the left part there is the Images table, which displays all project images andmethod of their processing, sensor, off-nadir angle, GCP count and information aboutadjustment steps;

54


The system allows to sort images by the column headings of the table and multiple selectionof pictures, using Ctrl and Shift keys.

• the Image search section allows to perform the search in the Images table. To dothis perform the following:

○ input the search query in the Image name field;

○ [optional] click the Search up button to set the list search direction;

○ [optional] click the Search down button to set the list search direction;

• the Parameters section allows to setup adjustment parameters, which set dependson selected method of images processing.

If the Import adjustment option is selected, there are no parameters, since all exterior ori-entation elements are imported from ADS 40/80/100 data, acquired after preliminary pro-cessing and adjustment of images block in dedicated software like ORIMA.

The Apply and Apply to all buttons allows to apply setup parameters to images, selectedin the Images table, or to all images of a block correspondingly.

• the Method section allows to select adjustment method:

○ Rigorous method – see Section 8.3.2;

○ RPC – method which uses RPC-coefficients – see Section 8.3.3;

○ Generic method – see Section 8.3.4;

○ Import adjustment – import of exterior orientation parameters from metadata.

The Import adjustment parameter is used only for ADS 40/80/100 projects.

• the Computing device section allows to select a computing device that will be usedfor adjustment procedure: CPU orGPU.

If working station is equipped with NVIDIA CUDA processor, choose theGPU option to speedup adjustment procedure (see detailed information on web-site http://www.nvidia.ru/ob-ject/cuda_gpus_ru.html).

In case of GPUwas chosen as an accounting device, but it initialize was failed or it has notenough of computational capability, CPUis used automatically.

• the Image groups section allows to setup images groups, acquired from the sameorbit:

55


http://www.nvidia.ru/object/cuda_gpus_ru.html

http://www.nvidia.ru/object/cuda_gpus_ru.html

If groups of images are available for block adjustment this allows to use less number of GCPper block (see recommendations in Section 8.6).

○ [optional] No groups – no image grouping;

○ [optional] For manual grouping of images that are used for adjustment procedure,select the User-defined groups option. Images groups are used only for imagesacquired from the same orbit.

○ [optional] For automatic creating of images groups by ties with automatic filteringof incorrect ties, select the Auto groups parameter.

In order to display information about filtered ties in a report, select Detailed report optionin the Image groups section of the Report tab.

• the Stereoscopic processing section allows to set theCreate stereopairs checkboxto perform adjustment with stereopairs creation.

○ [optional] In order to create stereopairs from every two adjacent images of a strip,select the Create stereopairs by strips option.

The Create stereopairs by strips parameter is used for stereoblocks only.

○ [optional] In order to form stereopairs from block images manually select the Ex-tended stereopair creating capabilities option and click the Stereopairs editor...button.

In order to calculate epipolar images using tie points, set the Use adjustment resultto create epipolar images checkbox off. Otherwise, the systemwill calculate epipolarimages just after adjustment, since the checkbox is set by default.

It is recommended to set the Use adjustment result to create epipolar images checkboxto obtain correct models on stereopairs with complicated enough geometry (at asynchronousstereo survey).

• theAdjustment steps section is used to perform step-by-step adjustment – success-ive adjustment of selected subblocks of scanner space images.

• the Enable optimization checkbox allows to significantly increase the system pro-cessing performance.

It is not recommended to clear this checkbox.

56


8.3.2. Rigorous method of adjustment

When perform adjustment by rigorousmethod the system considers exterior orientationparameters from images metadata, obtained from RSD provider.

Fig. 26. Parameters of rigorous method of adjustment

The rigorous method provides two ways of adjustment:

• Perform independent refinement – allows to adjust block images individually (alongwith applying of polynomial corrections), i.e. each image is adjusted considering justGCP measured on the image, without using ties with other block images;

Independent refinement is applied to obtain initial approximation for block adjustment or isused as a main operation of single image orientation (in case of excluding this image fromblock adjustment).

If the Separate position and attitude refinement checkbox is set the system appliescorrections to linear and angular exterior orientation parameters by turns.

57


• Involve into block adjustment (by default) – allows to adjust the whole images blockusing both GCP and tie points along with applying of polynomial refinements (undercondition that the Use ties in block adjustment checkbox is set on the Points tab).

To perform correct adjustment by rigorous method it is recommended to use default settings.

During adjustment procedure the system applies corrections to linear and angular ex-terior orientation parameters, obtained from images metadata. The corrections arepolynomials, which orders are calculated by the system automatically or are setup byuser manually. The corrections are setup depending on type of exterior orientationparameters:

• for linear exterior orientation parameters – Along track, Across track, Radial;

• for angular exterior orientation parameters – Roll, Yaw, Pitch.

It is recommended to use automatic corrections applying first, and then after obtainingadjustment results (during analyzing results shown in report or analyzing residualsvectors in the Images and Stereopairs modes) it is possible to make a decision whetherto setup refinement values manually for some exterior orientation parameters. In orderto setup the refinement manually select the User defined option in the Refinementpolynomial orders section, specify a polynomial order or set the – (it is set on by default)if it is necessary to not to submit a refinement.

Table 2. Interpretation of error vectors caused by inaccuracy of exterior orientation parameters

Direction of error vectorRefinement typeAlong track/Pitch

Across track/Roll

Radial

Yaw

58


8.3.3. RPC method

When perform adjustment by RPC method the system uses RPC coefficients frommetadata, obtained from RSD provider.

Fig. 27. Parameters of images adjustment using RPC coefficients

TheRefinement section allows to select one of the following types of corrections, whichare applied to RPC coefficients during adjustment:

• Auto – a correction type is defined automatically by number of GCP and tie points;

• Bias – constant correction will be applied to RPC;

• Affine – affine correction will be applied to RPC;

• Scale+bias – two constant corrections will be applied to RPC;

• None – no correction will be applied to initial RPC; this parameters is used for ex-ample, to check correctness of GCP or project coordinate system.

59


Parameters of the Parameters a priory covariances section allow to specify accuracyof initial RPC and accuracy of coordinate system specifying (for example, in case ofapproximation of local coordinate system with unknown parameters using topocentrichorizontal coordinate system).

Fig. 28. Parameters a priory covariances

The Parameters a priory covariances section contains the following parameters:

• Bias (pixels) – RMS of absolute terms of input refinement (in pixels) a priory;

• Drift (p.p.m.) – RMS of linear coefficients of input refinement a priory; it is dimension-less units and their value is setup in millionth parts (p.p.m. parts per million).

The system provides opportunity to use DEM for adjustment by RPC-method to increaseadjustment accuracy. There are the following recommendations concerning use of DEMduring adjustment:

1. It is recommended to use DEM in projects with relatively smooth relief, i.e. withsmooth elevation difference. Otherwise, sharp elevation difference in tie point areaor artefact on DEM lead to incorrect data.

2. It is not recommended to use DEM if there are a lot of triangulation points, since itresults in significant slowing down of adjustment procedure.

3. For stable work it is recommended to decrease the Bias parameter value on theParameters apriory covariances section in relation to default value.

Optimal value could be selected experimentally and could be both 1 and up to 0.01.

8.3.4. Generic adjustment method

Generic method supposes using affine, and parallel-perspective model or DLT (DirectLinear Transformation) algorithm and allows to process any scanner images (includingIRS, LandSat, etc.). However, at that it is necessary to have more GCP per stereopairas compared to rigorous algorithm (at least 4 GCP; recommended 10 GCP, see Sec-tion 8.6).

60


For orthomosaic processing (for example, to process images acquired by SPOT-5 with 2Аprocessing level, see more details in description of satellite system) only generic adjustmentmethod is used.

Fig. 29. Parameters of generic method of images adjustment

The Model type section is used to select one of the following model types dependingon a number of used GCP:

• Parallel-perspective – at least 7 GCP;

• Direct Linear Transformation (DLT) – at least 6 GCP;

• Affine – at least 4 GCP.

Themore the number of GCP the better the accuracy of adjustment results, however, whereverpossible it is recommended to use Rigorous or RPC-method, which provide better accuracyas compared to generic method results.

61


http://www.racurs.ru/wiki/index.php/SPOT-5

8.3.5. Manual creation of stereopairs

The system provides opportunity to form stereopairs from block images manually duringadjustment of images block.

Stereopairs forming is performed in the Stereopair editor window, which allows tocreate stereopairs automatically from adjacent images of each strip. In addition, it isalso possible to set stereopairs formed by non-adjacent images of a strip and to addinterstrip stereopairs. Besides, manual stereopairs forming allows to build epipoles forselected images.

To open a Stereopair editor window perform the following:

1. Click the button of the Block Adjustment toolbar. The Parameters windowopens;

2. Click the Images tab;

3. In Stereoscopic processing section set the Create stereopairs checkbox on.Select the Extended stereopair creating capabilities option;

4. Click the Stereopair editor button. The Stereopair editor window opens.

Manual creation of stereopairs is performed in those cases when it is necessary to formstereopair from images with small overlap area or with intersection angle in range from25 to 40 degrees.

62


Fig. 30. The “Stereopair editor” window

The Stereopair editor window contains the following sections:

• Project images – a table with all block images with specified imaging date and time,strip number and whether the image is included to a stereopair;

• Stereopair – allows to select two images from the Project images table to form astereopair from them;

• Selected image’s stereopairs – a table displaying all defined stereopairs, to whichis included the image, selected from the Project images table;

• Stereopairs list – list of all formed stereopairs.

To edit a list of created stereopairs use theDelete selected and theDelete all buttons, locatedunder the Selected image’s stereopairs and Stereopairs list tables.

The system provides the following methods of stereopairs generation:

• Create by strips – automatic stereopairs creation from adjacent images of eachstrip;

63


• Create by overlaps – allows to form stereopairs by parameters of images overlapin a strip;

• Create stereopairs manually.

To perform automatic stereopairs creation from adjacent images of each strip click theCreate by strips button. The stereopairs would be created in automatic mode.

To form stereopairs by parameters of images overlap in a strip, perform the followingactions:

1. Click the Create by overlaps button. The Find stereopairs by overlaps windowopens.

Fig. 31. Parameters of stereopairs search by overlaps

64


2. Set one or multiple checkboxes to search for maximal number of stereopairs usingselected parameters depending on the following tasks:

• to check correctness of stereopair forming from different time images or fromimages acquired from different sensors set the There are points measured onboth images;

• to form stereopairs in order to create digital elevation model set theOverlappingarea percentage checkbox on and specify overlap value in the At least field inpercent and/or set the Convergence angle checkbox on and specify a range ofangle values to the from..to fields in degrees;

• to limit stereopairs forming using number of pixels, hitting into images overlap,set the Overlapping area pixels count checkbox on and specify the At leastvalue in millions of pixels;

• to form stereopairs just from images acquired by the same sensors, set the Bothimages were acquired with the same satellite/sensor checkbox on;

• to search for images using restriction by survey data, set the Time gap checkboxon and specify a number of days in the At least... and Doesn’t exc...days fields;

• to search for images taken in the specified season, set the Imaging date differ-ence (disregarding years) checkbox on and specify a number of days in theDoesn’t exceed...days field;

• to limit stereopairs forming using value of spatial resolution, set the GSD differ-ence should not exceed checkbox on and specify the value in the field in percent.

• The Sun azimuth difference does not exceed checkbox sets maximum differ-ence of Sun position angles at the moment of taking each of the two images ofa stereopair;

• The Sun elevation difference does not exceed checkbox sets maximum differ-ence of the Sun elevation over the horizon at the moment of taking each of thetwo images of a stereopair;

In case when some of the parameters cannot be calculated or recognized on stereopairs(the If can’t be estimated option is used) the system provides the following two ways.Select the Reject in drop-down list to ignore stereopairs with mismatched parameters, orselect Accept to create stereopairs anyway. To reset all the parameters click the Resetparameters button.

3. Click the Find stereopairs button. In Stereopairs found table all the stereopairsthat match the search criteria are displayed as well as their properties.

To edit a list of created stereopairs use the Delete selected and theDelete all buttons.

65


4. Click OK to return to the Stereopair editor window. The stereopairs from Stereo-pairs found table would be automatically added to the Stereopairs list in Stereo-pair editor window.

To form stereopairs manually perform the following actions:

1. Select left image from the Project images table and click the Add image 1 buttonin the Stereopair section;

2. Select right image from the Project images table and click theAdd image 2 buttonin the Stereopair section;

3. Click the Create button to confirm and add stereopair to the Stereopairs list table.Otherwise click the Clear button;

4. [optional] Repeat these actions to create more stereopairs.

OK button in Stereopair editorwindow allows to create stereopairs for their further usein the adjustment procedure. Otherwise click the Cancel button.

8.3.6. Groups of images acquired from the same orbit

The system provides opportunity to create group of images, acquired from the sameorbit, that allows to use less GCP for scanner block adjustment. It is enough to use 4GCP per each image group. The system provides possibility to create images groupsautomatically (by sensor type, survey time and off-nadir angle), as well as possibilityto select images and make groups manually.

For automatic creation of images groups select the Auto groups option. The systemautomatically selects images acquired by the same sensor, and with the same surveytime and off-nadir angle.

To create groups from images acquired from the same orbit perform the following actions:

1. Analyze images suitability to be grouped: compare them by survey time (time differ-ence should not exceed 1 minute) and check images joint in overlap areas in theAerial triangulation module.

2. Measure tie points coordinates in overlap areas in order to group images. For correctimages grouping it is recommended to measure at least 2-3 points coordinates ineach images overlap.

When it is impossible to create images groups, it is necessary to increase the number oftie points, and the progress window displays the message about this error during imagesgroups creating (after pressing the button).

Tie points used for creating images groups cannot be involved to block adjustment.

66


measurement.pdf


4. On the Images tab set the User-defined groups checkbox on and click the Setgroups... button. The Image groups editor window opens.

Fig. 32. The “Image groups editor” window

5. [optional] If it is possible to group images from each strip, click the by projectstrips button. In order to create images groups using metadata click the bymetadata button.

To group images by strips it is necessary prior to groups creating split images into stripsby orbits in the block editor.

6. Set a name of new images group in the input field, located under the Image groupstable and click the Create button. A new image group will be created in the table.

7. In order to add images to created images group (to define images list to be grouped),select images in the Untied images table and click the button. Selected imageswill be shown in the Group images table. Statistic data about images group will beopened in a row of the Image groups table: number of images selected for groupcreation and imaging time difference of these images (minimal and maximal time,duration in seconds).

If imaging time difference is significant (more than one minute) creation of images groupbecomes doubtful. In this case it is recommended to perform additional research of imagesoverlap areas in the Aerial triangulation module (see the “Aerial triangulation” UserManual).

8. Create other images groups from the rest of images shown in the Untied imagestable.

67


measurement.pdf

9. Click OK to return to the Images tab of the Parameters window. Images groupswill be displayed in the Images table.

10. On the Report tab in the Image groups section setup display parameters ofgrouping results in a report, and specify image group join threshold.

It is recommended to set the RMS value less than 1 pixel. For images acquired from thesame orbit image group join threshold should not exceed 0.01 pixel.

11. Click the Compute button.

12. Click the Report button to view images grouping results in the Image groupssection and analyze errors values whether they exceed the threshold specified onthe Report tab of the Parameters window.

Tie points used for creating images groups cannot be involved to block adjustment.

Fig. 33. Brief report on images groups

68


Fig. 34. Detailed report on images groups

When it is impossible to create images groups, RMS considerably exceeds athreshold value, and the report displays a message about error during imagesgroups creating.

69


Fig. 35. Report on residuals obtained during images grouping

8.4. The “Report” tab

TheReport tab of the Parameterswindow is used to setup display of adjustment resultsand their level of details in the report, as well as to define acceptable residuals duringblock adjustment.

70


Fig. 36. Parameters on the “Report” tab

The Report tab contains the following parameters to be setup:

• the Tie points errors section is used to choose one of the following methods of re-siduals calculation for tie points:

○ From adjusted location – is used to calculate residuals by maximal deviation frommean coordinates value, calculated for all stereopairs;

○ Between stereopairs – is used to calculate residuals by maximal differencebetween values of tie point coordinates, calculated on different stereopairs.

• the Include into report section is used to setup display adjustment results in a report:

○ Report on images – is used to include adjustment results by images into report;

○ Report on stereopairs – is used to include a list of stereopairs and values of ste-reoangles (convergent angles) to report;

71


A list of stereopairs is displayed in report only when the adjustment is performed withstereopairs creation, i.e. The Stereopairs checkbox is set on the Images tab of theParameters window.

○ Report on block – is used to include adjustment results for the whole block to thereport;

○ Mark exceeding of limits – is used to mark residuals which exceed thresholdspecified in the Acceptable residuals section;

Residuals which exceed threshold are marked by the $QUOTE$" * $QUOTE$" symbol.

○ Exclude acceptable – is used to exclude from the report an information about re-siduals within the threshold, specified in the Acceptable residuals section;

○ Points catalogue – is used to include to the report a list of adjusted coordinatesof triangulation points (X, Y, Z);

○ Show errors – is used to add XY residuals values (Exy) and Z residuals values(Ez) to a list of adjusted coordinates;

○ Include orientation data – is used to add to the report coefficients of transfer toimage coordinates and coefficients refinements.

○ Include the list of points rejected by convergence angle – the list of points re-jected by convergence angle is added to the report.

• the Image groups section is used to setup display image groups results in a report,which includes images acquired from the same orbit, as well as to specify imagegroup join threshold (RMS):

○ Detailed report – is used to include full statistics (detailed and brief) about imagesgroups to the report;

○ Short report – is used to include only brief statistics about images groups to thereport;

○ Do not include in report – is used to exclude information about images groupsfrom the report;

○ Mark exceeding of limits – is used to detect and mark image grouping errors –RMS values that exceed threshold specified in the Image group join thresholdinput field;

○ Image group join threshold – is used to input RMS of images group thresholdvalue.

72


• theAcceptable residuals section is used to input residuals threshold values: controlpoints, check points and tie points on image and on ground;

• the Sort by section is used to setup display of a list of erroneous points in report:

○ point name – is used to display erroneous points in alphabetical order;

○ point error – is used to display adjustment errors in ascending order.

Errors in meters on scanner stereo block – is an error of intersection solution: X, Y, Z coordinatesare calculated by stereopair after adjustment and compared with X, Y, Z coordinates, specifiedin catalogue.

Errors in pixels – is an error of resection solution: using known ground coordinates (X, Y, Z) thecoordinates on image are calculated and compared with coordinates measured on image.During scanner block adjustment, in case of stereopairs absence, these errors for selected GCPor check point are shown in the Image residuals section on the Information window of theImages tab.

Error on image in meters – is a shift of the point on terrain, which is the distance between thepoint measured on the ground and the point, calculated taking into account the exterior orientationparameters.

8.5. Recommendations on monoblock adjustment

Monoblock – is a set of overlapped images, which usually have small overlap area.Most often a monoblock consists of images acquired from different sensors and withsignificant time interval between surveys.

When processing blocks containing images with significant survey time interval, workof correlator during automatic measurement of tie points is hampered. It takes placedue to following reasons:

• different luminosity parameters;

• survey in different seasons;

• different images resolution;

• change of situation on terrain (for example, construction of new objects, forest cutting);

• appearance of oblique images since off-nadir angle differs greatly for different blockimages;

• data relative accuracy depends on absolute accuracy of satellite position on the orbit.

There are the following recommendations for monoblocks processing on different stepsof adjustment:

73


1. On block forming step images are distributed to strips depending on a set of initialdata.

Fig. 37. Scheme of monoblock forming in the system

2. On step of triangulation points measurement there are the following recommenda-tions concerning setup of parameters of automatic tie points measurements:

74


Fig. 38. Parameters of automatic correlation for monoblock

3. On the step of mono block adjustment perform the following actions:

It is recommended to perform the mono block adjustment using the method of successiveapproximations with filtering.

1. Click the button of the Block adjustment toolbar. The Parameters windowopens.

2. On the Points tab input the following parameters:

• in the Points coordinates accuracy section input 0.5 pix to the Imagemeasurements RMS input field;

• in the Unity-weighted geodetic coordinates accuracy section input 1 mto the plane and height input fields;

75


• in the Points usage section input 0.01 to the Coordinates delta threshold(pix) input field;

• in the DTM for adjustment section set the Use DEM option and load theDEM from file.

During mono block adjustment it is highly recommended to use DEM due to insuf-ficient images overlap area. Otherwise, there are errors in the correlator work, aswell as errors in calculating Z coordinates of tie points measurements.

3. Click OK.

4. Click the button of the Block adjustment toolbar.

5. Tie points with tie residual that exceeds 10-20 pixels transfer into Excludedstatus in the By ties section of the Information window (see Section 5.5).

6. Set the Use ties in block adjustment checkbox in the Points usage sectionon the Points tab of the Parameters window.

For images with ’low-quality’ exterior orientation parameters it is recommended to setthe Affine option in the Parameters window on the Images tab in the Refinementsection.

For images with precise exterior orientation parameters it is recommended to set theNone in the Parameters on the Images tab in the Refinement section. In this casethe refinement will be not calculated.

7. Click the button of the Block adjustment toolbar to adjust the block oncemore.

8. Tie points with tie residual that exceeds 2-5 pixels transfer into Excluded statusin the By ties section of the Information window (see Section 5.5).

9. Perform adjustment accuracy control (see Section 11).

All steps of adjustment procedure require evaluation of adjustment results accuracy.Here the tie points measurements and exterior orientation parameters consistencycould be evaluated. During monoblock adjustment, the adjusted estimate of the accur-acy of tie points coordinates in meters is displayed as 0. This is due to the fact thatDEM serves as a reference data for the tie points. Therefore, quality control ofmonoblock processing reduces to estimating errors on tie points.

76


8.6. Number of points recommended for adjustment of scannerblock

There are some recommendations about number of GCP and check points used fordifferent methods adjustment of images included to scanner mono/stereo blocks.

For satellite scanner images it is possible to use only XYZ GCP (with X, Y, Z coordinates).

Fig. 39. Number of points recommended for adjustment of scanner block by rigorous method

If a monoblock contains images with triple or quadruple overlap area, it is recommended tomeasure several points in this area.

Number of GCP could be decreased due to tie points measured in triple or quadruple overlap.

77


Fig. 40. Number of points recommended for adjustment of scanner block by RPC method

Number of GCP could be decreased due to tie points measured in quadruple overlap.

78


Fig. 41. Number of points recommended for adjustment of scanner block by generic method

Recommended number of check points on images is not absolute and in many cases is determ-ined by the customer requirements to specific project. Nevertheless, it is necessary to havecheck points available for objective estimation of adjustment accuracy.

9. Adjustment procedureIn order to perform adjustment, it is necessary to setup parameters of project adjustmentand click the Compute button. The calculation procedure is started and the windowopens where the adjustment operation status is displayed.

79


Fig. 42. Window of adjustment procedure status

Adjustment operation convergence is represented by the diagram. The window alsoshows iteration number, adjustment method, and a value of refinement.

If the iterative process does not converge, interrupt the adjustment operation using theStop button to display a result obtained on current iteration. In order to delete all theadjustment results click the Cancel button.

Non-convergent adjustment may be caused by relative orientation errors, errors in GCP coordin-ates or by incorrect adjustment parameters (see Section 13), or due to wrongly defined coordinatesystem (right-left).

Adjustment procedure interrupts only between iterations, at that the adjustment results aresaved.

In order to save satisfactory adjustment results click the button of the Block adjust-ment toolbar. Pass to further project processing (see the “Project creation” UserManual).

The button on the Adjustment toolbar is used to restore current block status. Alsoit allows to load adjustment results after project loading.

When adjustment brings unsatisfactory results and it is necessary to bring the blockback to its initial status (as it was prior to adjustment launch), click the button on theAdjustment toolbar.

10. Creating adjustment reportThe system provides possibility to view full statistics of adjustment, summary informationabout adjustment residuals and control data.

80


project.pdf

To display adjustment report the button on the Block adjustment toolbar.

Fig. 43. The window of adjustment report

The toolbar of the Report window contains the following buttons:

• – allows to return to previous view of the report;

• – allows to refresh the report;

• – allows to start search in report’s text (Ctrl+F);

• – allows to save a report into text file;

The report also saves in the \backup project folder automatically with the adjustment results(only for central projection images).

• – allows to save a report into system resources;

• – allows to print a report.

The adjustment report displays the following values:

• residual values calculated on GCP, check, tie points, and projection centers for thewhole block and for each point;

• catalogue of points coordinates;

81


• images exterior orientation parameters;

• GPS corrections;

• corrections for exterior orientation parameters.

Report content is setup on the Report tab of the Parameters window (see Section 8.4 andSection 7.4).

Residuals and errors are displayed in the report as follows:

• X, Y, Z – coordinates values taken from model (from strip when using independentstrips method, from stereopair when using independent stereopairs method);

• X1, X2, Y1, Y2, Z1, Z2 – values of points coordinates on two different models;

• Xср, Yср, Zср – values of points coordinates, averaged over all models;

• Xg, Yg, Zg – geodetic value of points coordinates, specified by the user for GCP andcheck points;

• Exy, Ez – average residuals of points XY and Z coordinates;

• dX, dY, dZ, dS – discrepancies on GCP;

• dX, dY, dZ, dXY – GPS corrections on projection centers.

11. Adjustment accuracy controlIt is recommended to perform control of block adjustment accuracy after the completionof the adjustment operation. The adjustment results are displayed in graphic form as aresiduals vectors in the Stereopairs and Images modes of working windows, numericvalues of residuals are included to adjustment report.

If the adjustment brought unsatisfactory results and there some blunders the systemprovides possibility to improve the results using editing of points position. To do this itis necessary “re-measure” points, on which the residuals vectors exceed definedthreshold and are shown by red color.

In order to change point’s type or position perform the actions described above (seeSection 5.5).

Click the button on the Block adjustment toolbar to complete changing point’s po-sition on image.

In order to use interstrip points and points non-transferred in some strip in the adjustment byindependent strips method or independent stereopairs method, it is necessary to measure co-ordinates of the points on adjacent images inside the strip during aerial triangulation or in thePoints measurement window (see the “Aerial triangulation” User Manual).

82


12. Data export and import

12.1. Export of initial data

The system provides opportunity to perform export of source data for further adjustmentin third party software. The system allows to export data to AeroSys software format,as well as to PAT-B format.

Export or import of adjustment data is only possible in projects of central projection images.

For source data export perform the following actions:

1. Click the button on theAdjustment toolbar. The Export optionswindow opens.

Fig. 44. Export parameters

2. In the Export formats section set one or multiple formats checkboxes, to which itis necessary to export the data:

• AeroSys project – is used to perform further adjustment in theAeroSysprogram.Files of this format has the *.aer extension;

During export to AeroSys format all measurements of the system project are exportedto AeroSys project, see the official web-site of AeroSyscompany.

83


http://www.aerosys.com/

• The PAT-B (point coordinates XYZ) format – is used to perform further adjust-ment in third party software. File of this format contains coordinates of measuredpoints in meters. Files of this format has the *.xyz extension;

• The PAT-B (exterior orientation parameters – XYZ and matrix) format – isused to export coordinates of images projection and matrix of angular exteriororientation parameters. Files of this format has the *.ori extension;

• Exterior orientation parameters – XYZ and angles – is used to export coordin-ates of images projection and angular exterior orientation parameters. Files ofthis format has the *.cri extension;

• Images measurements – is used to perform further adjustment in third partysoftware. File of this format contains coordinates of measured points in pixels.Files of this format has the *.im extension.

3. [optional] In order to swap coordinate system (left/right) during adjustment import,in the Coordinates section set the Swap X Y checkbox.

If the Swap X Y checkbox is off, the initial data will be imported in the right coordinatesystem. Otherwise, in the left one.

4. Click the button. The Destination file window opens.

5. Choose a folder to place a file inWindowsfile system.

6. Specify file name where to export adjustment results.

When input a name of some file the same file name and path is set for all selected formatsautomatically.

7. [optional] To cancel automatic specifying of the same name for all selected formats,set the Files autonaming checkbox off.

8. [optional] To correct an accuracy of GCP geodetic coordinates, projection centersand exterior orientation parameters, edit weight values of these parameters in theWeights section (default value is 1). In output file a value of accuracy will be dividedby specified weight.

9. [optional] To measure a priory RMS accuracy of points measurements on images,set the image measurements value in mm in the Measurements accuracy(std.dev.) section.

This is used for less quality source images. By default it is set 0.5 pixel (0.005 mm) valuefor measurements, obtained from scanned images acquired by precisely calibrated camera.

84


Besides measurements (points coordinates on images, geodetic coordinates of GCP’s,exterior orientation parameters), the a priory accuracy of these measurements is writtento the exported files as well.

10. [optional] In order to export adjustment results in necessary measurement units (inmicrons, for example), in the Image measurements scaling input field input acoefficient which all measurements will be multiplied by.

11. Click OK to complete the export operation.

12.2. Import of adjustment results

The system provides opportunity to perform import of adjustment results obtained usingthird party software.

Export or import of adjustment data is only possible in projects of central projection images.

For import of adjustment results obtained in the AeroSys software or other software packages,it is recommended to choose Cartesian coordinate system for the system project, otherwise,XY and Z residuals may occur in coordinates of images projection centers.

To import adjustment results, obtained in the AeroSys software perform the followingactions:

1. Click the button on the Adjustment toolbar. The Adjustment import windowopens.

Fig. 45. Parameters of import from AeroSys project

2. In the Import format section set the AeroSys project format.

3. Click the button. The Open AeroSys project window opens.

85


4. Specify a name and path of the AeroSys project in Windows file system (file with*.aer extension).

5. [optional] In order to change coordinate system during the adjustment import, inthe Coordinates section set the Swap XY checkbox.


6. Click Import to complete import operation from AeroSys project to the system.

To import adjustment results from file, which contains coordinates of images projectioncenters and matrix of angular exterior orientation parameters, perform the followingactions:


Fig. 46. Parameters of import from PAT-B format

2. In the Import format section set the PAT-B (rotation matrix) format.

3. In the PAT-B – point coordinates XYZ field click the button. The Open PAT-B file – point coordinates window is opened.

4. Specify a name and path of the file with points coordinates in Windows file system(file with *.xyz extension).

5. In the PAT-B – exterior orientation parameters – XYZ and matrix field click thebutton. The Open PAT-B file – exterior orientation parameters window

opens.

6. Specify a name and path of the file with points coordinates in Windows file system(file with *.ori extension).

86




8. [optional] In order to change a direction of coordinate system if the coordinate sys-tems directions are mismatched, in the Angles section set the Negate focuscheckbox on.

9. [optional] In order to invert a matrix of angles rotation, set the Invert matrixcheckbox on in the Angles section.

10. Click the Import button to complete import from PAT-B format to the system.

To import adjustment results from file, which contains coordinates of images projectioncenters and exterior orientation angles, perform the following actions:


Fig. 47. Parameters of import from PAT-B format

2. In the Import format section select the PAT-B (angles) format.

3. In the PAT-B – point coordinates XYZ field click the button. The Open PAT-B file – point coordinates window is opened.

4. Specify a name and path of the file with points coordinates in Windows file system(file with *.xyz extension).

5. In theExterior orientation parameters – XYZ and angles field click the button.TheOpen exterior orientation parameters file – XYZ and angleswindow opens.

87


6. Specify a name and path of the file with points coordinates in Windows file system(file with *.сri extension).



8. [optional] In order to change a direction of coordinate system if the coordinate sys-tems directions are mismatched, in the Angles section set the Negate focuscheckbox.

9. [optional] In order to invert a matrix of angles rotation, set the Invert matrixcheckbox in the Angles section.

10. Click the Import button to complete import from PAT-B format to the system.

13. Stages of accuracy control

13.1. Adjustment control in free model

13.1.1. Splitting on subblocks

Splitting on subblocks is a necessary condition for adjustment control operation in freemodel in the following cases:

• if the block contains many blunders, it is easy to detect them by splitting on sub-blocksin such a way that each sub-block would contain not more than one blunder;

• if a processed block has a length more than 20 bases, time for sub-block adjustmentis less than time for adjustment of the whole block.

13.1.2. Blunders control in camera focal length

In order to detect blunders in camera focal length value, it is recommended to input asurvey basis approximately. If a basis value was input correctly, the Scale section onthe Report tab of the Parameters window displays correct value of survey scale afteradjustment. If the difference between calculated height values of triangulation pointsand projection centers differs significantly from real ones at that, than more likely focallength was input incorrectly.

During adjustment in free model values of calculated Z-coordinates of projection centers areclose to 0.

During adjustment in free model values of calculated Z-coordinates of triangulation points arenegative and close to survey altitude by absolute value.

88


During adjustment in free model small errors in value of camera focal length cannot be detected.In this case it is recommended to perform adjustment using GCP.

13.1.3. Blunders control in strips tying

In order to detect blunders in interstrip tie measurements it is recommended to adjustsub-blocks using independent strips method in free model.

In case of long strips (20 bases) the errors might exceed calculated ones in 10 20 times.

It is recommended to split on parts strips with length more than 10 20 bases.

In order to detect errors within a threshold, it is necessary to specify a survey basis withaccuracy of 30-50%. Operation of checking correctness of specified basis is performedafter adjustment in the Parameters window on the Report tab in the Scale section bycalculated value of survey scale.

In order to obtain correct error values (in meters), it is necessary to specify a surveybasis (with accuracy of 30-50%). Correctness of specified basis checks after adjustmentin the Parameters window on the Report tab in the Scale section by calculated valueof survey scale.

For blocks with strips of moderate length (up to 15 images), it is recommended to adjustthe whole block using independent strips method in free model.

13.1.4. Intermediate control of tie measurements blunders

After control of adjustment that was performed by independent strips method in freemodel, it is recommended to execute control tie points measurement blunders usingadjustment by independent stereopairs method. The errors that are not exceeding 2 4thresholds are acceptable.

13.1.5. Blunders detection on interstrip not-transferred points

In order to detect blunders in interstrip points, non-transferred in images overlap on oneof the strips, it is recommended to perform block adjustment using independent stereo-pairs method in free model (prior to bundle adjustment).

“Not-transferred” point – points which coordinates are measured only on a single image in onestrip (it is not measured in overlap of adjacent images in a strip).

If point’s coordinates are measured in adjacent images overlap area, it is necessary tocheck tie points distribution in triplets corresponding to this point. If the ties are locatedin some small area of overlap, it is recommended to measure additional tie points onthe whole overlap area.

89


13.1.6. Control of errors of tie points measurements

In order to detect blunders in interstrip and inside strip tie points measurements, it isrecommended to perform block adjustment in free model using independent stereopairsmethod, then adjustment the block by bundle adjustment method. It is recommendedto alternate two ways of adjustment until obtaining error values, that not exceed 75%of tie errors threshold before executing adjustment with GCP, since using GCP dataleads to increasing of error values.

The first adjustment should be performed by independent stereopairs method, as it ismore stable. Then it is possible to alternate independent stereopairswith bundle adjust-ment.

To obtain correct error values of interstrip and inside strip tie points measurements thatwithin threshold, it is recommended to set survey basis approximately with 10–20% ofaccuracy.

13.1.7. Control of errors of automatic tie points measurements

During automatic tie points measurement it is possible to face a situation when obviouslyerroneous points are corrected, and maximal (out of tolerance) adjustment error existson points measured correctly. One of the reasons may be incorrect work of automatictie points measurement algorithm (due to complex terrain, wrong algorithm settings).

In order to detect errors of automatic tie points measurement it is recommended tocheck regularity of points’ distribution and errors absence on closest stereopairs andtriplets.

It is also recommended to perform control of each point, starting from points with max-imal error. Perform the following actions to do this:

1. Check point location (measurement) in stereomode:

• point should not be placed in trees shadow;

• correctness of point position in Z-axis.

2. Check point recognition correctness:

1. Measure one more point near point under test (on a distance not exceeding ofabout 1/10 of stereopair overlap area).

2. After adjustment check error value on both points:

• if the errors on both points are of the same value (direction and size), thenthe point is recognized correctly;

90


• if error on the tested point is within tolerance threshold, and error on a newpoint has maximal value, it is necessary to perform control of points regulardistribution;

• if the errors on both points have a value close to accuracy limit, it is recom-mended to complete point’s validation.

3. Check a few closest tie interstrip points (all of them or minimum 3 ones regularlydistributed in each stereopair overlap where the point exists, and somemore pointsin a distance of 1 3 stereopairs).

4. Check errors on new tie points:

• measure one tie point in those triplets where the tested point is located;

• check errors on new tie points after adjustment operation. The error will appearon really erroneous points.

If it is impossible to identify erroneous points after all the procedures, it is necessary tocheck other possible sources of errors:

• interior orientation errors;

• deformations of some particular image of the block;

• incorrect choice of objects for points’ measurements (see Section 14.7).

To obtain tie points errors that are within a tolerance threshold, it is recommended todelete points with maximal error in adjacent stereopairs and to re-measure points inthis area manually.

13.2. Adjustment control with geodetic reference

13.2.1. Control of coordinate system correctness

If the adjustment of georeferenced block fails, it is necessary to check coordinate systemselection.

In order to detect coordinate system selection, perform the following actions:

1. Check if it is correct selection of left-handed or right-handed coordinate system:

• adjust block in the Cartesian right-handed coordinate system;

• adjust block in the Cartesian left-handed coordinate system;

• compare the adjustment results and check for residuals.

91


2. Selecting of the coordinate reference system type window opens. Check correctnessof a type of used coordinate system by comparing coordinates of 3 GCP, that areclearly not lie along a straight line, with their location on block scheme before ad-justment.

3. Make sure that there are no input mistakes in coordinates – presence of typos incoordinates (extra digits, missed digits, incorrect zone number).

13.2.2. Control of adjustment performed using projection centers co-ordinates

In order to define errors of adjustment with specified projection centers coordinates,perform the following actions:

1. Select GCP with a blunder on a block scheme.

2. In the Point attributes window click the button, to change selected GCP intocheck point.

3. Click the button on theBlock adjustment toolbar to perform a block adjustment.

4. Check for the presence of the random component of the error.

If there is a big constant Z-error in a block after adjustment, it is recommended to checkcamera focal length and coordinate system selection.

If there is a big systematic Z-error in a block after adjustment, Z-error vectors in a blockcenter are directed to opposite side in relation to vectors located on the edges of ablock. In this case it is recommended to check coordinate system selection (to considerthe curvature of the Earth) and check interior orientation (to consider distortion).

13.2.3. Control of adjustment performed without projection centers

If there is big enough amount of GCP (approximately 10 and more) on a block, it is re-commended to set small weight of equation on GCP (0.01–0.1) and adjust the blockwithout projection centers.

If there is a big systematic Z-error in a block after adjustment, Z-error vectors in a blockcenter are directed to opposite side in relation to vectors located on the edges of ablock. In this case it is recommended to check coordinate system selection (to considerthe curvature of the Earth) and check interior orientation (to consider distortion).

If there is a big systematic XY-error in a block after adjustment, vectors of XY errorsare directed radially from block center – the block is compressed in one direction andstretching of the block in perpendicular direction. In this case it is recommended tocheck for stretching/compression of images in one direction.

92


If the adjustment was performed by independent strips and independent stereopairs method,when there was set small weight of equations on GCP (approximately up to 0.01 and less), theblock will be compressed.

If the block was adjusted by bundle adjustment method, it will be not compressed.

13.2.4. Control of bundle adjustment errors

When perform adjustment by bundle adjustment method or import of adjustment resultsfrom external software, it is recommended to set in the Parameters window on theReport tab:

• in the Tie residuals section set the from mean value checkbox;

• in the Include in the report section it is recommended to set the Residuals, onstereopairs, include singles checkboxes.

If bundle adjustmentmethod was used to adjust block of images, acquired by long-focuscamera or which have images overlap coefficient that exceeds 65%, then when the focallength or images overlap coefficient increase there is a reducing of survey basis/heightratio, in that way the Z-errors on GCP are reduced.

When the adjustment was performed by the independent stereopairs method, so reducing ofsurvey basis/height ratio results in increasing of Z-errors on GCP.

When perform adjustment by bundle adjustment method Z-error on GCP transformsinto Z-error on tie points. In the report Z-error on GCP is displayed in the Tie point re-siduals (from mean value) section.

When perform adjustment by bundle adjustment method error on tie points, which co-ordinates are measured on a single stereopair, in some cases is defined by verticalparallax residual and difference of distances to projection centers of two stereopair im-ages on tie points. In the absence of adjustment blunders, the tie points residual tendsto zero and is negligible.

14. Special aspects of airborne blocks adjustment in thesystem

14.1. Use the coordinates of projection centers

During adjustment with specified projection centers coordinates it is recommended touse Cartesian coordinate system (for example, WGS-84).

When perform adjustment with specified projection centers coordinates it is not recommendedto use local coordinate system, which takes into account the curvature of the Earth approximately.

93


During adjustment without projection centers using local coordinate system it is not re-commended to consider as precise images exterior orientation parameters obtainedby the adjustment. When using a local coordinate system occurs a discrepancy betweenthe Z and XY coordinates of the projection centers, as well as between the angularelements of exterior orientation.

14.2. Survey with big images overlap

The main feature of a survey with big images overlap (more than 70%) is that surveybasis to height ratio is small for stereopairs, composed by neighbor images of strip.Therefore, the value of height error on points measured on these stereopairs exceedsplane errors in several times. However, at that 100% of block area is covered by overlapsof three and even more images of a strip. In stereopairs composed by images takenas next nearest (one per 2, 3, depending on overlap size), the survey basis to heightratio is big enough like during usual survey with 60% overlap, that allows to obtain smallheight error.

The best way of processing of such images blocks depends on kind of output production.

If the end production is orthomosaic creation using exterior DEM, perform the followingactions:

1. Perform the whole block processing.

2. Measure coordinates of each point (tie, GCP, check) on all images of a strip, wherethis point is displayed.

3. Perform adjustment using bundle adjustment method.

4. In adjustment report it is recommended to consider only errors on images and errorson GCP and check points shown in the Images mode of working window.

If it is necessary to produce 3D vector objects or DEM, it is recommended to reducenumber of images in such a way that images overlap in resulting block will be close to60%, and then process this block in a usual way. It is recommended to perform suchimages reducing when images overlap is not less than 75 80%.

In order to evaluate images reducing coefficient, use the following expression:

(40 / (100−k))−1, where k – images overlap in percent.

Analyze the results:

• final value equals 0 – it is recommended to process all images in strips;

• final value equals 1 – prior processing it is recommended to reduce number of imagesand involve next second image;

94


• final value equals 2 – prior processing it is recommended to reduce number of imagesand involve next third image;

In order to create orthomosaic using external DEM and 3D vector objects or DEMschoose one of the following ways:

• prior to block processing reduce number of images in strips in such a way that imagesoverlap in resulting block will be about 60%;

• split a block into two subblocks (full and reduced). Then adjust full block and importthe adjustment results into the reduced block. Use the full block for orthomosaiccreation, and the second one for building 3D vector objects or DEMs.

14.3. Survey by long-focus cameras

The main feature of survey performed by long-focus cameras is that resulting imagesare rectangles with considerable difference between width and height.

To obtain 3D vector objects or DEMs it is not recommended to use survey performed by long-focus cameras, since considerable height errors occur during the project processing.

In order to decrease elevation errors on images when perform survey by long-focuscamera, perform the following actions:

• when possible, turn a camera in such a way, that image long side will be orientedalong flight;

• specify images overlap equal 60%; in this case survey basis-height ratio will bemaximal and therefore height errors during block adjustment will be minimal;

• place and measure interstrip ties in “zigzag” line, if used geodetic network is notdense enough.

14.4. Survey in different seasons

During processing of images block which contains multi-temporal images, includingthose acquired in different seasons, there is a problem of tie points measurement.

If images strips were acquired in different seasons, then during tie points measurementin image areas, where interstrip stereo visualization is impossible, it is recommendedto measure a tie point on the same object individually in each strip. When measuringtie points on the ground level it is necessary to consider a refinement on height differenceof grass cover, which covers an object footing.

In images strips acquired in different seasons, it is recommended to measure tie pointson artificial objects (power lines pillars, buildings, roads).

95


14.5. Survey of forested areas

During processing of images block which contains huge forested areas there is aproblem of tie points measurement.

When processing such a survey a residual of neighboring strips exceeds an error thatcaused by incorrect point positioning. Adjustment accuracy in areas with continuousforest cover and in neighboring areas may be impaired.

If it is impossible to exclude densely forested areas from images block processing, it isrecommended to measure tie points on artificial objects (power lines pillars, buildings,roads).

14.6. Bundle adjustment

During bundle adjustment operation in the Stereopairs mode of working window occursdeviation from “mean” adjusted position of interstrip ties points by elevation on differentstereopairs to the same side. Height deviation is determined by the X-parallax on inter-strip tie points located in stereopairs formed by images of different strips.

During adjustment using independent stereopairs method a deviation from “mean” adjustedposition by elevation doesn’t occur.

14.7. Tie points measurement on lengthy objects boundaries

If the boundary or its vicinity has no any other objects that are located along theboundary, that are crossing the boundary, common irregular background or boundarycurvature, it is not recommended to measure a tie point in the boundary vicinity.

Lengthy objects boundaries – boundaries of objects and objects themselves: roads (tracks),arable lands (consisting of parallel furrows) and other similar objects.

Appendix A. Exterior orientation angular systemsThe angular exterior orientation parameters are calculated and included to the reportas ’omega-phi-kappa’ or ’alpha-omega-kappa’ system of angles.

Rotation sequence in ’omega-phi-kappa’ angular system.

a)

96


Fig. A.1. Rotation about X-axis at a positive value of “omega” angle

b)

Fig. A.2. Rotation about Y-axis at a positive value of “phi” angle

c)

Fig. A.3. Rotation about Z-axis at a positive value of “kappa” angle

Rotation sequence in ’alpha-omega-kappa’ angular system.

a)

97


Fig. A.4. Rotation about Y-axis at a negative value of “alpha” angle

b)

Fig. A.5. Rotation about X-axis at a positive value of “omega” angle

c)

Fig. A.6. Rotation about Z-axis at a positive value of “kappa” angle

98


Documents

USER MANUAL - Racurs13.1.6.Controloferrorsoftiepointsmeasurements.....89 13.1.7.Controloferrorsofautomatictiepointsmeasurements.....90 1.Purposeofthedocument